Form DOT F 1700 · KAB Crashes 4824 39.6 44,606 25.9 Intersection 1734 14.2 41,112 23.8 Intersection-Related 1331 10.9 32,798 19.0 Driveway Access Related 1092 9.0 16,296 9.4 Non-Intersection

Technical Report Documentation Page 1. Report No.

FHWA/TX-02/4048-1

2. Government Accession No.

3. Recipient's Catalog No.

5. Report Date

October 2001

4. Title and Subtitle

CHARACTERISTICS OF AND POTENTIAL TREATMENTS FOR CRASHES ON LOW-VOLUME, RURAL TWO-LANE HIGHWAYS IN TEXAS

6. Performing Organization Code

7. Author(s)

Kay Fitzpatrick, Angelia H. Parham, Marcus A. Brewer, and Shaw-Pin Miaou

8. Performing Organization Report No.

Report 4048-1

10. Work Unit No. (TRAIS)

9. Performing Organization Name and Address

Texas Transportation Institute The Texas A&M University System College Station, Texas 77843-3135

11. Contract or Grant No.

Project No. 0-4048 13. Type of Report and Period Covered

Research: September 2000-October 2001

12. Sponsoring Agency Name and Address

Texas Department of Transportation Research and Technology Implementation Office P. O. Box 5080 Austin, Texas 78763-5080

14. Sponsoring Agency Code

15. Supplementary Notes

Research performed in cooperation with the Texas Department of Transportation and the U.S. Department of Transportation, Federal Highway Administration. Research Project Title: Low-Cost Design Safety Improvements for Rural Highways 16. Abstract

Low-volume, rural two-lane highways carry less than 8 percent of the total vehicle miles on state-maintained (or on-system) highways and have approximately 11 percent of the total on-system vehicle crashes. These roadways also have relatively more severe injuries when vehicle crashes do occur. For example in 1999, about 26 percent of the Texas on-system crashes were KAB crashes (i.e., fatal, incapacitating injury, or non-incapacitating injury crashes), while over 40 percent of the crashes on low-volume on-system roads in 1999 were KAB crashes. In general, crashes on low-volume, rural two-lane highways occur between intersections by a single vehicle running off the road. Crashes on curves and in dark, non-lighted conditions are more common on low-volume, rural two-lane highways than on urban roads. The study in crash characteristics also demonstrated that more KAB crashes occurred in eastern counties than western counties in Texas. A sample of counties was selected to investigate which regional characteristics are associated with high and low crash rates. In general, sites in the eastern counties had less driver-friendly characteristics—more horizontal and vertical curves, narrower lanes and/or shoulders, less forgiving roadside development, higher access density, and higher roadside development scores. Eastern counties also had more crashes at intersections than western counties. To obtain information about the types of treatments being used on these types of facilities, a mailout survey was distributed to the TxDOT districts and to other states. A literature review was conducted to identify the known effectiveness of treatments used on rural two-lane highways. 17. Key Words

Low-Volume (<2000 ADT), Rural, Two-Lane Highways, Crashes, Treatments

18. Distribution Statement

No restrictions. This document is available to the public through NTIS: National Technical Information Service 5285 Port Royal Road Springfield, Virginia 22161

19. Security Classif.(of this report)

Unclassified

20. Security Classif.(of this page)

Unclassified

21. No. of Pages

242

22. Price

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

CHARACTERISTICS OF AND POTENTIAL TREATMENTS

FOR CRASHES ON LOW-VOLUME, RURAL TWO-LANE HIGHWAYS IN TEXAS

by

Kay Fitzpatrick, Ph.D., P.E. Research Engineer

Texas Transportation Institute

Angelia H. Parham, P.E. Assistant Research Engineer


Marcus A. Brewer, E.I.T. Assistant Transportation Researcher


and

Shaw-Pin Miaou, Ph.D. Research Scientist


Report 4048-1 Project Number 0-4048

Research Project Title: Low-Cost Design Safety Improvements for Rural Highways

Sponsored by the Texas Department of Transportation

In Cooperation with the U.S. Department of Transportation Federal Highway Administration

October 2001

TEXAS TRANSPORTATION INSTITUTE

The Texas A&M University System College Station, Texas 77843-3135

v

DISCLAIMER

The contents of this report reflect the views of the authors, who are responsible for the opinions, findings, and conclusions presented herein. The contents do not necessarily reflect the official views or policies of the Texas Department of Transportation (TxDOT) or the Federal Highway Administration (FHWA). This report does not constitute a standard, specification, or regulation, nor is it intended for construction, bidding, or permit purposes. The report was prepared by Kay Fitzpatrick, P.E. (TX-86762), Angelia Parham, P.E. (TX-87210), Marcus A. Brewer, and Shaw-Pin Miaou. The engineer in charge of the project was Kay Fitzpatrick.

vi

ACKNOWLEDGMENTS The research team recognizes Danny Brown, the project director; Lynn Passmore, the program coordinator; and technical panel members Robert Neel, Margaret (Meg) Moore, Aurora (Rory) Meza, and David Bartz for their time in providing direction and comments for this study. The research team would also like to recognize the more than 125 individuals who provided information on treatments by responding to our mailout survey. We also would like to thank the following district representatives for meeting with members of the research team during interviews conducted as part of the research: Robert Neel, Herbert Binkley, Richard Ivey, Carlos Chavez, Edgar Fino, Patricia Dalbin, Matt Carr, Daniel L. Dalager, and Imelda Barrett. The research team would also like to express appreciation to the following Texas Transportation Institute staff who assisted with data collection and with developing materials for this report: Maria Medrano, Stephanie Elmquist, Justin Burns, Jeremy Davis, Charles Stevens, Todd Hausman, and Dan Walker. The research reported herein was performed by the Texas Transportation Institute as part of a study entitled “Low-Cost Design Safety Improvements for Rural Highways” and was sponsored by the Texas Department of Transportation in cooperation with the U.S. Department of Transportation, Federal Highway Administration.

vii

TABLE OF CONTENTS CHAPTER 1 INTRODUCTION...................................................................................................1-1 2 FINDINGS FROM SURVEYS AND INTERVIEWS ...........................................2-1 MAILOUT SURVEY ....................................................................................2-1 INTERVIEWS WITH DISTRICT REPRESENTATIVES .........................2-23 3 LITERATURE REVIEW........................................................................................3-1 OTHER VALUABLE REFERENCES..........................................................3-1 4 VEHICLE CRASHES ON ON-SYSTEM, LOW-VOLUME, RURAL

TWO-LANE HIGHWAYS IN TEXAS..................................................................4-1 STATE-LEVEL ANALYSIS.........................................................................4-2 DISTRICT-LEVEL ANALYSIS ...................................................................4-5 COUNTY-LEVEL ANALYSIS ..................................................................4-10 SUMMARY OF FINDINGS .......................................................................4-15 5 EVALUATION OF CRASHES AT ROADWAY LEVEL....................................5-1 SITE SELECTION.........................................................................................5-1 DATA COLLECTION...................................................................................5-3 DATA ANALYSIS ........................................................................................5-5 SUMMARY OF FINDINGS .......................................................................5-14 6 SUMMARY AND CONCLUSIONS .....................................................................6-1 REFERENCES....................................................................................................... R-1 APPENDICES A LANE DEPARTURE.............................................................................................A-1 B ROADWAY DEPARTURE .................................................................................. B-1 C ROADWAY CROSS SECTION ........................................................................... C-1 D ALIGNMENT ........................................................................................................D-1 E ROAD SURFACE CONDITION .......................................................................... E-1 F NARROW BRIDGE ...............................................................................................F-1 G INTERSECTIONS.................................................................................................G-1 H WILDLIFE CRASHES..........................................................................................H-1 I WORK ZONES........................................................................................................I-1 J CRASHES BY AREA/VOLUME .......................................................................... J-1 K CRASHES BY DISTRICT GROUP......................................................................K-1

1-1

ADT = 2000 to 60007% ADT > 6000

14%

ADT > 600018%

ADT < 20001%

ADT < 200030%

Rural

Urban

ADT = 2000 to 600030%

CHAPTER 1

INTRODUCTION

The state of Texas maintains nearly 80,000 centerline-miles of paved roadways serving about400 million vehicle miles per day. Over 62 percent of the centerline-miles are rural two-laneroads that, on average, have less than 2000 ADT (average daily traffic). These low-volume ruralroadways carry less than 8 percent of the total vehicle miles on state-maintained (or on-system)highways but have approximately 11 percent of the total on-system vehicle crashes. When onlytwo-lane highways are considered, almost three-fourths of the crashes occur in the ruralenvironment with 30 percent of the crashes occurring on the low-volume roads (see Figure 1-1).

Due to the low volume and relatively low crash frequency on these roads, it is often not cost-effective to upgrade the roads. However, vehicles traveling on these roadways generally havehigh speeds and, thus, tend to have relatively more severe injuries when vehicle crashes do occur. For example in 1999, about 26 percent of the Texas on-system crashes are KAB crashes (i.e.,fatal, incapacitating injury, or non-incapacitating injury crashes), while over 40 percent of thecrashes on low-volume on-system roads in 1999 were KAB crashes (See Table 1-1).

Figure 1-1. Distribution of Crashes by ADT on Two-Lane Highways in Texas.

1-2

Table 1-1. Low-Volume (< 2000 ADT), Rural Two-Lane Highway Crashes for 1999.

On-System, Low-Volume, Rural Two-Lane Highway Crashes

All On-System Crashes

Frequency Percent Frequency Frequency

PDO: Non-Injury 4407 36.2 58,288 33.8

C Crashes: Possible Injury 2959 24.3 69,836 40.4

B Crashes: Non-Incapacitating 2946 24.2 31,902 18.5

A Crashes: Incapacitating Injury 1418 11.6 10,331 6.0

K Crashes: Fatal 460 3.8 2373 1.4

TOTAL 12,190 100.0 172,730 100.0

KAB Crashes 4824 39.6 44,606 25.9

Intersection 1734 14.2 41,112 23.8

Intersection-Related 1331 10.9 32,798 19.0

Driveway Access Related 1092 9.0 16,296 9.4

Non-Intersection 8033 65.9 82,524 47.8

TOTAL 12,190 100.0 172,730 100

Little information exists to help transportation practitioners evaluate the effectiveness of low-costmeasures, especially the effectiveness on low-volume roads. Therefore, there is a need toprovide information that discusses the safety improvement options for low-volume roadways. Objectives for year one of a Texas Department of Transportation project included identifyingcommon types of crashes on low-volume roadways; characteristics of low-volume, rural two-lanehighway crashes; and potential safety improvements. This report summarizes the project’s first-year activities.

This report is divided into the following chapters and appendices:

• Chapter 1 contains an introduction concerning crashes on low-volume, rural two-lanehighways.

• Chapter 2 provides information gathered from a mailout survey and from interviewsconducted at four TxDOT district offices.

• Chapter 3 introduces the methodology used to conduct the literature review for the project,the appendices that contain the details from the review, and summaries of other referencesthat could be valuable when selecting treatments for a low-volume, rural two-lane highway.

• Chapter 4 presents information on vehicle crashes for on-system, low-volume, rural two-lanehighways in Texas. It provides answers to three questions: how often do crashes occur,where do crashes occur, and what types of crashes occur more often.

• Chapter 5 discusses an evaluation of the differences in crashes between counties in theeastern and western portions of the state.

1-3

• Chapter 6 summarizes the findings from the year one efforts of the TxDOT project.• Appendix A provides information on treatments for lane departure crashes.• Appendix B discusses treatments for hazards located in the roadside.• Appendix C presents suggestions on treatments that are within the roadway cross section.• Appendix D reviews treatments used along an alignment.• Appendix E discusses treatments used to decrease crashes associated with wet pavement.• Appendix F provides information on treatments for narrow bridges.• Appendix G presents treatments for crashes at intersections or driveways.• Appendix H provides information on animal crashes and treatments that have been used.• Appendix I provides an overview of sources of information on treatments for work zone

crashes.• Appendix J presents the details on the statistical characteristics of vehicle crashes for three

ADT groups within rural and urban environments.• Appendix K presents the statistical characteristics of vehicle crashes for three district groups.

2-1

CHAPTER 2

FINDINGS FROM SURVEYS AND INTERVIEWS

MAILOUT SURVEY

A mailout survey was conducted to gather information on relatively low-cost safetyimprovements on low-volume roads. (For purposes of this project, low-volume roadways aredefined as two-lane roads with an ADT �2000.)

A total of 98 surveys were mailed to: all 25 district engineers in the state of Texas (with copies toforward to the area engineers in each district); district engineers (or the equivalent) in the statesof California, Florida, and Washington; and one design engineer in each of the remaining states.Respondents were asked to:

• check those safety improvements they have installed to address safety concerns on low-volume two-lane roads (by checking the items on the list provided);

• list the three to five most recent safety improvements used on a low-volume two-lane road toaddress a safety concern;

• list candidate sites for the improvements listed in the first question (asked of Texasrespondents only);

• provide additional comments or suggestions; and• indicate if they would like to receive a copy of the survey results.

Texas produced 75 responses while other states offered 49 responses. The following pagessummarize the 124 survey responses received. Eighty-nine respondents asked to receive a copyof the survey results, while 18 respondents did not want to receive a copy, and 23 respondentsdid not respond to this question. A one-page, front-and-back summary was prepared anddistributed to those requesting a copy of the survey results.

QUESTION 1. Please check the safety improvements you have installed to address safetyconcerns on low-volume, two-lane roads (ADT �2000).

This question was divided into eight categories with several safety improvements listed in eachcategory. The differences between the responses from Texas and other states are summarized ingraphs with supporting text. For the graphs, the values on the vertical axis indicate thepercentage of responses for that question. For example, in Clear Zone Improvements, 75 percentof the Texas respondents have removed trees to improve the clear zone, while 87 percent ofrespondents from other states have removed trees to improve the clear zone. Table 2-1 lists thenumber and percent of responses.

2-2

Table 2-1. Installed Safety Improvements.

Potential Safety TreatmentsTexas Responses Other State Responses

Number Percent (%) Number Percent (%)

Clear Zone

Flatten side slopes 50 66 37 79

Increase clear zone 43 57 35 75

Make culverts traversable by adding bars to preventtires from entering culvert

60 79 14 30

Mow 62 82 29 62

Remove headwalls or adding fill to bring groundlevel with headwall

53 70 27 57

Remove trees 57 75 41 87

Upgrade safety appurtenances 64 84 42 89

Other 7 9 5 11

Wildlife Control

Methods to control wildlife management 2 3 6 13

Reflectors to alert wildlife of approaching vehicles 0 0 9 19

Sign (with or without flashers) to alert drivers ofwildlife

39 51 39 83

Other 1 1 1 2

Additional Lane

Climbing lane 13 17 22 47

Passing lane 14 18 16 34

Right-turn lane 39 51 26 55

Left-turn lane 42 55 24 51

Two-way left-turn lane 21 28 14 30

Other 2 3 1 2

2-3

Table 2-1. Installed Safety Improvements (continued).



Pavement Surface Treatments

Centerline rumble strips 0 0 10 21

Edgeline rumble strips 5 7 19 40

Rumble strips on approaches to intersections orhorizontal curves

10 13 23 49

Shoulder texturing 8 11 5 11

Skid resistance improvements 41 54 23 49

Thicker thermoplastic pavement markings 38 50 8 17

Other 1 1 1 2

Pavement Markings

Add on-lane pavement markings (painted curvearrow, slow speeds, etc.)

14 18 9 19

Add oversized glass beads 22 29 8 17

Add pavement markings (e.g., edgelines) 46 61 34 72

Add raised pavement marker on centerline oredgeline

57 75 24 51

Add retroreflective pavement markers 21 28 13 28

Reapply existing pavement markings because theyhave faded

49 65 35 75

Remove existing buttons to convert to guidancemarkings

28 37 5 11

Other 1 1 1 2

2-4

Table 2-1. Installed Safety Improvements (continued).



Sign Improvements

Advance signing for intersections 51 67 44 94

Advance signing for horizontal curves 57 75 41 87

Advance signing for stop signs 64 84 39 83

Delineators 60 79 40 85

Diamond grade sheeting at restricted width bridge 26 34 17 36

Diamond grade chevron signs at curves 27 36 24 53

Flags on stop sign 17 22 6 13

Flashing beacon on stop sign 24 32 21 45

Flashing beacon on warning sign 32 42 19 40

High intensity strobe (HIS) in advance of curves 8 11 2 4

In-rail reflectors for guardrail and bridge rail 31 41 29 62

Reflective corner caps on signs (contrasting colors) 9 12 1 2

Other 1 1 1 2

Signal Improvements

Backboards for traffic signals 15 20 11 23

High intensity strobe (HIS) in signal 7 9 5 11

Other (please list): 1 1 0 0

Other Improvements

Illumination 20 26 21 45

Improve/standardize approaches to narrow bridges 29 38 16 34

Increase pavement edge maintenance 50 66 33 70

Speed detection/notification devices 17 22 5 10

Other (please list): 2 3 2 4

2-5

0

10

20

30

40

50

60

70

80

90

100

Sid

e Sl

opes

Cle

ar Z

one

Cul

verts

Tra

vers

able

Mow

Rem

ove

Hea

dwal

ls

Rem

ove

Tree

sS

afet

y Ap

purte

nanc

es

Oth

er

Per

cen

t

Texas

Other States

Clear Zone Improvements

Upgrading safety appurtenances, removing trees, mowing, flattening side slopes, removing oradding fill around headwalls, and increasing clear zone had high responses from Texas and fromother states (see Figure 2-1). One difference between the two groups was that 79 percent ofTexas respondents said they had made culverts traversable, while only 30 percent of other staterespondents checked this item.

The “Other” responses to this category included adding shoulders, moving metal beam guardfence further from the edgeline, providing safety lighting at intersections, trimming trees andbrush, closing drainage to eliminate ditch lines, utility pole relocation, delineation of trees andutility poles, removing fixed objects, improving access location and sight distance, and addingguardrail.

Figure 2-1. Clear Zone Improvements.

2-6

0

10

20

30

40

50

60

70

80

90

100

Con

trol M

etho

ds

Ref

lect

ors

Sign

s

Oth

er

Per

cen

t

Texas

Other States

Wildlife Control

Signs to alert drivers of wildlife are used widely in other states (83 percent), while only 51percent of the Texas responses indicated that signs are used (see Figure 2-2). Also, 19 percent ofother states use reflectors to alert wildlife of approaching vehicles, and none of the Texasrespondents reported using this measure.

The “Other” responses to this category included adding culvert crossings as well as providinghorse and duck crossings.

Figure 2-2. Wildlife Control.

2-7

0

10

20

30

40

50

60

70

80

90

100

Clim

bing

Lan

e

Pass

ing

Lane

Rig

ht-T

urn

Lane

Left-

Turn

Lan

eTw

o-W

ay L

eft-T

urn

Lane

Oth

er

Per

cen

t

Texas

Other States

Additional Lane Improvements

Texas and other states’ responses for the use of left-turn lanes, right-turn lanes, and two-way left-turn lanes were very similar (see Figure 2-3). However, other states use climbing lanes (47percent versus 17 percent) and passing lanes (34 percent versus 18 percent) more frequently thanTexas respondents.

The “Other” responses to this category included: providing deceleration lanes at private driveswith high ADTs (i.e., plants and stockyards); adding wider shoulders where driveways,mailboxes, or intersections are frequent enough that a large number of vehicles are entering orexiting the travel way; and using slow-moving vehicle turnouts in areas with poor passingopportunities and high recreational vehicle (RV) use.

Figure 2-3. Additional Lane Improvements.

2-8

0

10

20

30

40

50

60

70

80

90

100

Cen

terli

ne R

umbl

e St

rips

Edge

line

Rum

ble

Strip

sIn

ters

ectio

n R

umbl

e St

rips

Shou

lder

Tex

turin

g

Skid

Res

ista

nce

Impr

ovem

ents

Thic

ker M

arki

ngs

Oth

er

Per

cen

t

Texas

Other States

Pavement Surface Treatments

Texas and other state respondents indicated similar uses of skid resistance improvements andshoulder texturing (see Figure 2-4). However, Texas has a much higher usage of thickerthermoplastic pavement markings than other states (50 percent versus 17 percent). Other stateshad a much higher usage of centerline rumble strips (21 percent versus 0 percent), edgelinerumble strips (40 percent versus 7 percent), and rumble strips on approaches to intersections orhorizontal curves (49 percent versus 13 percent).

The “Other” responses to this category included using larger glass beads and paved shoulders.

Figure 2-4. Pavement Surface Treatments.

2-9

0

10

20

30

40

50

60

70

80

90

100

On-

Lane

PM

sO

vers

ized

Bea

ds

Add

PMs

Add

RPM

s/C

ente

r or E

dge

Add

RPM

sW

ider

Edg

elin

es

Rea

pply

PM

sR

emov

e Bu

ttons

Oth

er

Per

cen

t

Texas

Other States

Pavement Markings

Texas and other states listed similar uses for adding on-lane pavement markings (PM), addingedgelines, adding retroreflective pavement markings (RPM), and reapplying existing pavementmarkings because they have faded (see Figure 2-5). Other states’ respondents use wider edgelinemarkings more frequently than Texas respondents (19 percent versus 5 percent). Texasrespondents use three treatments more frequently than other state respondents: oversized glassbeads (29 percent versus 17 percent), raised pavement markers on centerlines or edgelines (75percent versus 51 percent), and removing existing buttons to convert to guidance markings (37percent versus 11 percent).

The “Other” responses to this category included using pavement marking rumble strips and usingedgeline striping regardless of the roadway width.

Figure 2-5. Pavement Markings.

2-10

0

10

20

30

40

50

60

70

80

90

100

Adva

nce

Inte

rsec

tions

Adva

nce

Cur

ves

Adva

nce

Stop

sD

elin

eato

rs

Dia

mon

d Sh

eetin

g Br

idge

Dia

mon

d C

hevr

ons

Flag

s on

Sto

ps

Beac

ons

on S

tops

Beac

ons

on W

arni

ngs

Stro

be C

urve

sIn

-rail

Ref

lect

ors

Ref

lect

ive

Cor

ner C

aps

Oth

er

Per

cen

t

Texas

Other States

Sign Improvements

Texas and other state respondents listed similar use of advance signing for horizontal curves,advance signing for stop signs, delineators, diamond grade sheeting at restricted width bridges,flashing beacons on stop signs, flashing beacons on warning signs, high intensity strobes inadvance of curves, and in-rail reflectors for guardrail and bridge rail (see Figure 2-6). Texasrespondents indicated more use of flags on stop signs than other state respondents (22 percentversus 13 percent) and of reflective corner caps of contrasting color on signs (12 percent versus 2percent). Other state respondents indicated more use of diamond grade chevron signs at curvesthan Texas respondents (53 percent versus 36 percent). The “Other” responses to this category included installing signs at intersections (W-10) andadding “orange mouse ears” on signs.

Figure 2-6. Sign Improvements.

2-11

0

10

20

30

40

50

60

70

80

90

100

Back

boar

ds

Stro

bes

Oth

er

Per

cen

t

Texas

Other States

Signal Improvements

Texas and other state respondents indicated similar uses of backboard for traffic signals and forhigh intensity strobes in traffic signals (see Figure 2-7).

The “Other” response to this category included replacing loops in the pavement with videodetectors.

Figure 2-7. Signal Improvements.

2-12

0

10

20

30

40

50

60

70

80

90

100

Illum

inat

ion

Brid

ge A

ppro

ache

sPa

vem

ent E

dge

Mai

nten

ance

Spee

d D

etec

tion

Dev

ices

Oth

er

Per

cen

t

Texas

Other States

Other Improvements

Texas and other states indicated similar uses of improving or standardizing approaches to narrowbridges and increasing pavement edge maintenance (see Figure 2-8). However, Texasrespondents indicated more use of speed detection and notification devices (22 percent versus 11percent), and other states indicated more use of illumination (45 percent versus 26 percent).

The “Other” response to this category included rumble strips, lane widening, guardrails, roadwaygeometry, and providing a 2-ft paved shoulder.

Figure 2-8. Other Improvements.

2-13

QUESTION 2. Please describe the 3 to 5 most recent safety improvements you have used(or plan to use) on a low-volume, two-lane road (ADT �2000) to address a safety concern.

The responses are summarized by the condition treated on the following pages. Because thenumber of responses for most items was low, the responses are presented in number of responsesrather than in percentages (as reported in Question 1). Respondents could check more thanone condition treated for each safety improvement.

A. Roadside Objects—The highest responses for Texas and other states for safetyimprovements associated with roadside objects were safety end treatments, sign improvements,increased clear zone, pavement resurfacing, removing or trimming trees, and adding shoulders(see Table 2-2).

Table 2-2. Number of Respondents Indicating Use of These Improvements Used as a Treatment for Roadside Objects.

Safety Improvement T* O* Safety Improvement T* O*

Add or improve safety end treatmentsSign improvements including sheeting, upgrading, posts, and one strobeIncrease clear zonePavement resurfacing or rehabilitation/ gradingRemove or trim trees / brushAdd shouldersAdd reflectors or delineators for guard railAdd guardrail or guard fenceWiden bridge or improve bridge approachPave or improve shouldersPavement edge maintenance or improvementMow, clear brush, or clear right-of-wayAlignment improvementsWiden roadwayAdd raised pavement markers on edgelinesRemove island and replace with stripingAdd climbing lanes

3913

99

775

44

43

3221

11

114

53

440

20

00

1200

01

Add raised medianAdd / widen pavement markings Upgrade mailboxesReconfigure intersectionFlatten slopesIntersection sight improvement or sight distance improvementIlluminationReplace / Improve guardrailConsistent lane and shoulder widthCorrect superelevation Right-turn taperShoulder texturingLedge removal to prevent rock from fallingEliminate ditchEdgeline rumble strips Truck escape rampUtility pole initiativeFlashing beacons on warning signsImprove intersection: remove island, reconfigure 4-way stop

11110

0000000

000000

0

02003

3111111

111111

1

*T - Number of Texas Responses*O - Number of Other State Responses

B. Driver Inattention—Highest Texas responses for safety improvements to reduce driverinattention were: sign upgrades; safety end treatments; raised pavement markers; advanceflashers; widening the road, shoulders, or lanes; and adding, improving, or reapplying pavementmarkings (see Table 2-3). The highest responses from other states included: chevrons ordelineators in curves; edgeline rumble strips or shoulder texturing; roadway realignment; sign

2-14

upgrades; widening the road, shoulders, or lanes; adding, improving, or reapplying pavementmarkings; and intersection rumble strips.

Table 2-3. Number of Respondents Indicating Use of These Improvements Used as a Treatment for Driver Inattention.


Sign upgrades and postsSafety end treatmentsInstall advance flashersInstall raised pavement markersWiden road, shoulders or lane, or add shouldersAdd, improve markings, or reapply pavementAdd delineators or reflectors on guardrailWiden bridgesPave shoulders or improve pavement edge maintenanceAdd oversized stop, or advanced stop signs and/or barsEdgeline rumble strips or shoulder texturingAdd turn lanes (left, right, or two-way left-turn)Flatten side slopesInstall skid resistant surfaceUpgrade 4-way flashersImprove clear zoneAdd bridge and guardrailAdd chevrons and delineators in curvesInstall raised median or center islandUpgrade mailboxInstall warning sign at T-intersectionIntersection rumble strips

128686

6

5

44

2

2

2

2222231111

43024

4

1

00

2

5

2

50032101004

Strobes in signal head / black signal facesUpgrade school flashers Flashing signalRoadway realignmentRemove / trim trees or brushThicker thermoplastic stripingGrade separation structurePassing laneIllumination ISD improvementMowTruck escape rampRumble stripsCenterline rumble stripsRumble strips on pedestrian pathRumble strips on curvesOversized speed limit signs Advance signs Intersection warning programCurve warning programFluorescent yellow cross-road signAdditional sidewalk, crosswalk, curb, cut ramps, and signsAdd centerline stripingOversized and high-intensity truck warning signsWarning signs for school bus stopTraffic signal

111111111000000000000

00

000

200620001312411112111

11

111


C. Shadows and Blinding—The highest Texas responses to reduce shadows and blinding wereremoving or trimming trees and bushes, installing flashing beacons on warning signs, and addingor reapplying pavement markings (see Table 2-4). There was only one response from anotherstate for shadows and blinding, and it included installing signs, advance road signs, oversizedtruck warning signs, safety end treatments, turn lanes, and through lanes.

2-15

Table 2-4. Number of Respondents Indicating Use of These Improvements as a Treatmentfor Shadows and Blinding.


Remove trees / trim brushInstall flashing beacons on warning signsImprove sight distanceAdd / reapply pavement markingsIlluminationMore delineation on guardrailsAdd rumble strips

32

12111

00

00000

Improve signsAdvance road signsOversized truck warning signsReplace / upgrade guardrailSafety end treatmentsAdd turn laneAdd through lane

0000000

1111111


D. Rural Intersections—The highest Texas responses for improving safety at rural intersectionswere upgrading and standardizing signs; installing left-turn lanes; advance warning flashers;rumble strips and advance signing; realign alignment; illumination; and adding pavementmarkings (see Table 2-5). The highest other state responses were chevrons and curve warningsigns; wildlife reflectors; resurfacing or adding chip seal; upgrading or standardizing signs;shoulder texturing or rumble strips; and improving clear zones.

Table 2-5. Number of Respondents Indicating Use of These Improvements Used as a Treatment for Rural Intersections.


Upgrade / standardize signsLeft-turn lane at intersectionInstall advance warning flashersAdd rumble strips and advance signingRoadway realignmentIlluminationAdd pavement markingsRight-turn laneUpgrade 4-way flashersAdd flashing beacons on stop signs or 4-way stop signsSafety end treatmentsReflective strips on stop signsShoulder texturing or shoulder rumble stripsWiden roadwayImprove clear zoneGrade separation structureAdd passing lane

5443333222

211

1111

2000110000

002

0200

Advance stop signs with flags at T- intersectionAdvance stop signs with flashing lights4-way stopWarning sign at T-intersectionStrobe in signal headAdd two-way left-turn laneBouncing lights and buttonsPainted warning on roadwayLarger stop signsCurve warning signs / chevronsWildlife reflectorsResurface / chip sealAdvance road signsInstall guardrailFlatten slopesIntersection warning programReplace guardrailImprove intersection sight distanceMow

1

111111110000000000

0

000010003321111111


2-16

E. Unexpected Alignment Changes—The highest Texas responses for improvements forunexpected alignment changes were raised pavement markers; chevrons, signs, and delineatorson horizontal curves; improving and upgrading signs; and adding pavement markings (see Table2-6). The highest responses from other states were chevrons, signs, and delineators on horizontalcurves; and safety end treatments.

Table 2-6. Number of Respondents Indicating Use of These Improvements Used as a Treatment for Unexpected Alignment Changes.


Raised pavement markersDelineators / chevrons / warning signs on horizontal curvesImprove / upgrade signsAdd pavement markingsSafety end treatmentsIn-rail reflectors on guardrailsFlashers on warning signsAdd 10-foot shouldersResurfaced roadwaySpeed detection and notification devicesClimbing lanesRaise headwalls to decrease slopesRumble strips (intersection and centerline)Reapply pavement markingsIncrease thickness of thermoplastic pavement markings

74

33222111111

11

18

01400010012

00

Delineation at bridge endsShoulder texturing / shoulder rumble stripsRealignment to reduce curve severityAdvance warning flashers for intersections and stop signsAdvance warning flashers on stop and stop ahead signsReplace bridge beamIlluminationImprove skid resistanceWildlife reflectorsUpgrade guardrailTruck escape rampCut trees / brushImprove clear zoneOGAC

11

11

1

111000000

01

10

0

001212211


F. Unexpected Developments (small towns, factories, etc.)—Only two respondents providedsafety improvements for unexpected developments—one from Texas and one from another state(see Table 2-7).

Table 2-7. Number of Respondents Indicating Use of These Improvements Used as a Treatment for Unexpected Developments.


Upgrade school flashersWiden roadwayImprove clear zoneSign standardizationSigns (watch for slow moving vehicles)

11111

00000

Curve warning signsMowExtend culvert headwallUtility pole initiativeAdd raised pavement markings

00000

11111


2-17

G. Weather—The highest number of safety treatments from Texas respondents for weatherconditions were safety end treatments, resurfacing or seal coating the roadway, raised pavementmarkers, shoulders, improving skid resistance, and removing trees or fixed objects from the clearzone (see Table 2-8). Other state respondents listed curve warning signs and/or chevrons andrumble strips most frequently, although the number of responses was low.

Table 2-8. Number of Respondents Indicating Use of These Improvements Used as a Treatment for Weather.


Safety end treatmentsResurface / seal coat roadwayRaised pavement markersAdd shoulders Improve skid resistanceRemove trees or fixed objects / clear zoneAdd pavement markingsIncrease thickness of thermoplastic pavement markingsImprove drainage (at intersection, at guardrail)Climbing lanesReapply existing pavement markingsUpgrade signing

88755

42

2

2111

01001

20

0

0001

Signs at intersectionsAdd edgelinesRaise grades in low areasPavement edge maintenanceMore delineation at guardrailsCurve warning signs / chevronsRumble stripsEdgeline rumble stripsIntersection rumble stripsFlatten slopesImprove guardrailFlashing beacon on warning signsFencingMinor sight benches

11111000000

000

00000321111

121


H. Wildlife Encroaching on Roadway—Only two survey respondents provided improvementsfor wildlife encroaching on the roadway, indicating no defined safety improvements for thiscondition (see Table 2-9).

Table 2-9. Number of Respondents Indicating Use of These Improvements Used as a Treatment for Wildlife Encroaching on a Roadway.


Widen roadwayImprove clear zoneWiden shoulders

111

000

Clear brushWildlife reflectorsSigns

100

100


2-18

I. Other Safety Concerns—Survey respondents were asked to list other safety concerns (notincluded in categories A through H) and to list the safety treatments installed in response to theseconcerns. The responses are summarized in Table 2-10.

Table 2-10. List of Other Safety Concerns and Safety Treatments Installed in Response.

Safety Concern Safety Treatment T* O*

Increase skid resistance • Surface friction improvement• Seal coat

22

0

Sight distance • Tree trimming or removal• Intersection warning program

2 1

Narrow roadway Widen road, resurface, safety end treatments, signs,minor alignment improvements

2 0

Narrow lanes Lane widening with maintenance operations 1 0

Passing opportunities /truck passing

Added passing lanes (Super 2 roadway) 3 0

Low shoulder or drop-off • 2-foot pavement widening on inside of curves• Closed drainage and eliminated ditch line• Add material to level surface• Widened edges on 2-lane highways to eliminate

drop-off and narrow lanes

11

11

Dead trees falling on road Tree removal 2 0

Edges of roadway Improved edges of roadway 1 0

Non-traversable ditchesand structures

Changed geometry of large roadside ditch blocks sovehicles involved in off-roadway excursions maytraverse the drainage structure

1 0

Clear zone • Removed metal beam guard fence (MBGF) andextended box culverts

• Safety end treatments

11

0

Durability Thermoplastic striping and raised pavement markers 1 0

Improve stopping sightdistance

Improve profile and skid resistance 1 0

No edgeline Widen roadway by adding 2-foot shoulder 1 0

Rutting Mill and inlay to remove rutting 1 0

Speed Advance stop signs, T-intersection, flags 0 0

Unsafe maneuvers bydrivers

No parking signs and channelizing devices 0 0

Errant traffic and saferecovery

Raise headwalls to decrease slopes 0 0

Grades Rebuilding FM roadways 0 0

2-19

Table 2-10. List of Other Safety Concerns and Safety Treatments Installed in Response (continued).

Safety Concern Safety Treatment T* O*

Passing Directional arrow and sign for headlight use 0 1

Run-off-road Shoulder rumble strips 0 1

Sub-standardsuperelevation

Correct superelevation 0 1

Drowsy drivers Edgeline rumble strips 0 1

Right-of-way violation Place stop signs on local street 0 1

Rocks falling onto road Ledge removal to prevent rocks from falling 0 1

Upgrade to currentstandards

• Replace single-lane bridge with 2-lane structure• Replace or upgrade guardrail and end anchorages• Extend culvert headwall to proper clear zone

requirement

000

111

Line of sight Slope flattening 0 1

Pedestrian safety Rumble strips along pedestrian path 0 1

High crash location Flashing beacons on warning signs 0 1

Driving Under theInfluence (DUI)

Corridor Review: Added ‘Buckle Up’ and ‘Alcohol.08 foot signs

0 1

Lighting Illumination 0 1

Sight distance Mow 0 1


QUESTION 3. Do you have sites that would be a good candidate for a treatment listedpreviously (Yes/No)?

This question was asked of Texas respondents in order to find candidate sites for this project. Of the 77 Texas responses, 38 respondents listed candidate sites, 19 did not have candidate sites,and 20 did not respond to this question.

QUESTION 4. Do you have any additional comments or suggestions?

Texas and other state respondent comments are listed and grouped by categories. The commentsor suggestions are printed in italics.

2-20

Correlation with Crash Data

In depth study of crash data, before and after “improvements” are completed to see where thereduction is: crash frequency or crash severity.

On roadways of this traffic volume, we have used many of these improvements but they are notusually applied until a situation where several crashes happen at a given time.

Guardrail, Bridges, and Shoulders

Guardrail upgrades and safety end treatment grades are needed.

We continue to add safety end treatments to culverts, widen bridges, and pave shoulders on U.S.and state highways.

Issues Are Addressed by Other Departments

Most work on low volume FM roadways has been isolated to the problem areas handled withmaintenance forces. Complete sections of roadways have been addressed in our hazardelimination program, but at this time, this work is concentrated to our higher volume roadwaysdue to budget constraints.

The Alpine area office considers safety improvements on all proposed projects at the designconcepts phase.

We do not do safety improvement projects as stand-alone projects in the Roadway DesignDivision. Traffic and Maintenance Divisions do more “safety improvements” projects; however,the Roadway Design Division includes the safety improvements marked on previous sheets in ourreconstruction, widen/overlay projects.

Safety improvement projects in the Northern Virginia District of VDOT tend to be located onhigh-volume urban roads. The purpose of our low-volume rural projects is generally to pavepoor quality gravel roads and bring them into overall compliance with state and AASHTOguidelines.

Improvements to less than 2000 ADT state roads are usually completed by force account andsomewhat difficult to track; many counties do their own signing, vertical realignment and otherimprovements.

In answering this survey, we assumed that the questions applied to “stand-alone” projects. Other safety improvements have been used as part of other, larger parties.

Work has been done by maintenance via traffic operations’ requests and help with funding.

2-21

Lighting

I have always thought it would be a good idea to illuminate rural “T” intersections. I believethis would help get the driver’s attention.

Pavement Surface

Our district added rumble strips to pavement at all stop intersections in rural areas.

Texturing/rumble strips are limited to four-lane facilities. We need something for two-lane roadswith comparable effects for errant/sleepy motorists–many fatalities.

Centerline rumble strips may be an effective, low-cost measure for improving the highway safety.VDOT has studied optimal shoulder rumble strips and implemented about 790 miles of it onVirginia Interstate system. We may study this later.

I have used centerline rumble strips on a two-lane pavement to reduce crashes. Three yearsprior, I had six fatal crashes with 12,000 ADT. Five years have passed and no fatal crashes. Thestrips covered about 2.5 miles of road at a cost of $13,000 and a 20,000 AADT today. I havemore information if you are interested.

Widening the shoulder is another useful improvement.

Pavement Width and/or Shoulders

Our biggest problems on low-volume roads are narrow pavement and edge drop off. Justimproving the roadway crown to 26 feet and using rip-rap to backfill along the edge of pavementis a big help.

Review advance signing for curves, intersections, etc., roadside delineation, minimum three-footpaved shoulder ribbons.

Policy / Procedure Issues

Concentrate on things that do not require ROW. May be more advantageous to widen base andsurface, move ditches out, leave slopes steep outside clear zone. ROW purchase on low-volumeroadway not high priority.

We are one of the smaller rural districts within Caftans, and although a majority of our roadsare rural two-lane, historically the highways with < 2000 ADT have not been safety problems.Our district policy is to place open graded asphalt cement and rumble strips when practical.

DOT tailgating safety initiative is based on the two-second rule to address aggressive drivingbehavior.

2-22

In general, our roadways with volumes under 2000 ADT don’t get a lot of attention unless theyshow up as a hazardous crash location or a risk location in our programming system. Theseroadways are typically bituminous surface treatments. When we apply a new chip seal, weupgrade guardrail ends and signing, and delineation. If there is a spot safety location with abenefit/cost ratio (b/c) over one we will fix it at the same time.

Signing, Reflectors, and Markers

I have several locations that have advance signing, but we still have problems with peoplerunning through the intersections.

Crash history is greater at horizontal and vertical curves. The addition of low-cost reflectivedevices helps guide traffic even when placed in locations not according to standards. Installation of chevrons on curves following standards forces large spacing between chevrons,and with the seven-foot height, the target value is not too great in dark, inclement weather.

We have one rural skewed intersection where we have utilized pavement markers to attempt topromote stopping traffic to stop perpendicular to the through traffic. According to localofficials, we believe this has decreased the crashes at this intersection.

Flashing beacon on warning sign and stop sign.

Our pavement marker (raised reflective) seems to receive a lot of positive feedback from thepublic – more than anything.

Slopes

Distances have become a major factor for this kind of road with AADT < 2000 in Virginia. Thestandard/criteria may need to be studied and reviewed.

Wildlife

Prairie dogs are doing damage to foundation of roadway; and we need to provide a way toprevent this legally.

Other

We do not have specific hazard-cause crashes that I have attempted to rectify. Items marked in 1are improvements added to roadways in construction/rehabilitation projects.

Add approach guardrail to bridge, safety dikes (escape ramps), and opposite “T” intersections;reduce horizontal degree of curvature; improve crest vertical; widen bridge; widen pavement;widen shoulders; relocate roadway to eliminate problems with horizontal/vertical curvature;eliminate bridge/culvert headwalls with new curvature meeting clear zone requirements; addchevrons; pave all part of shoulder; pave all or part of shoulder in curve and carrysuperelevation of curve into paved outside shoulder; and roll over into 6:1 forecloses.

2-23

This information relates to rural, non-state roads; and we are unable to provide specific locationinformation at this time due to the overall number of agencies (e.g., cities, towns, counties)involved.

INTERVIEWS WITH DISTRICT REPRESENTATIVES

Members of the research team also met with representatives of four districts along with gatheringinformation from the mailout survey. Meeting objectives included gathering information on howthe district identifies locations for treatments and how the treatments are selected. Table2-11 lists the questions used during the meetings. The four districts visited were Austin, El Paso,Lufkin, and Odessa. These are the districts responsible for the roads included in the evaluationof crashes on a selection of control-sections (see Chapter 5).

Table 2-11. Potential Questions for Meeting with District Representatives.

• Are low-volume rural roads treated differently than higher volume rural roads or urban facilities (e.g.,identification of sites, funding, type of treatments, etc.)?

• Which positions within your district have responsibilities for identifying safety needs and developing safetytreatments?

• Do you regularly conduct crash studies to identify high-crash locations?If so, how often?

• How do you decide which intersection to treat?• How do you identify potential countermeasures?• How often do you use a consultant to assist with this process?• In what areas do the consultants assist?

Data collectionsIdentify high crash locationsDeveloping recommendationsDeveloping design plansConstructing the improvements

• Do you have a hierarchy for safety improvements for an intersection? A roadway segment? What are thedifferent levels of improvements?

• Do you have an example of a site that has undergone several improvements in response to a safety issue? Ifso, please describe experience.

• Did the media or community requests play a role in the timing or types of treatments?• How do you fund safety improvements?• Do you use reference manuals when conducting a safety study?• What would you say is the most common improvement used within your district (e.g., signalization,

pavement markings, etc.)? What are the most effective?• Are there any potential countermeasures that you believe would be effective but haven’t tried yet? If yes,

what?• Do you have potential sites for before-and-after studies for countermeasures in addition to those provided in

the fall 2000 survey?

Key items from the meeting include the following:

• Each district participates in the Hazard Elimination (HES) program. The HES program ispart of the Highway Safety Improvement Program. The basic objective of the HES programis to reduce the number and severity of crashes. The districts prepare a Safety Evaluation

2-24

Report (SER) form for each proposed highway safety project. These forms are submitted inmid-November to the Traffic Operations Division who ranks the projects using the SafetyIndex and selects those approved for funding. In 2000, the funding level was approximately$36 million. The funds available within the HES program provide for the majority of thesafety treatments implemented within a district. Some districts mentioned that, in a fewcases, if a project was not funded through the HES program, they would use other funds totreat a location. Both rural and urban locations are considered within the HES program. Onerepresentative noted that rural two-lane low-volume roads may be at a disadvantage infunding competitions because the formula has ADT or axle as a variable.

• The Odessa District has a formal Safety Review Committee. This committee reviews everyfatal crash. As part of the review, they obtain information on other crashes at the site and visitthe site. The committee includes representatives of other public agencies such as theMetropolitan Planning Organization. They are encouraged to “think outside of the box”when identifying treatments. El Paso also mentioned their Safety Review Committee as amechanism for improving safety within their district. Their committee reviews plans forsafety concerns at 30, 60, and 90 percent completion on large projects and once on smallerprojects. Their meetings are scheduled on a project-specific basis.

• Potential locations are generally identified from either a district employee’s knowledge of theroadway system or from complaints made to an area office or the district. Locations arerarely identified by using the crash database to identify intersections or roadways with highcrash numbers or high crash rates. An exception to this is the annual wet weather review thatis performed to identify locations with a high number of wet weather-related crashes.

• Consultants are used to perform traffic counts, delay studies, identify treatments for a specificlocation, and develop the plans for a location. They are not used to identify high-crashlocations.

• Treatments for a site are determined either based upon an engineer’s judgment afterreviewing the crash pattern or within a brainstorming session of a safety review committee. The recommendations are reviewed by others within the department as plans are beingdeveloped or as the SERs are being completed. Sources for ideas on treatments include:previous experience within the district, treatments being used in other districts (either fromdriving in other districts or conversations at meetings like the Transportation Short Course),findings from research studies, and suggestions from vendors. For most districts, there doesnot appear to be one key reference being used to generate ideas. One district suggested thathaving the information on a website would be more valuable than within a printed document.

• All districts mentioned the increased use of video detection at signalized intersections. Thegeneral consensus is that it is better than in-pavement loops and that its use will continue toincrease.

• The most common types of improvements mentioned by the districts include:� signals,� safety end treatments,

2-25

� updating pavement markings (especially left-turn bays),� signs,� left-turn lanes (with pavement markings and signs),� increased shoulder width,� illumination, and� buttons or rumble strips (buttons are being used in some locations due to the rumble strip

depression being filled with sand or dirt that frequently blows in the area).

• Treatments being considered by districts include the following:� advanced rumble strips (also called audible strips),� shoulder texturing, and � rumble strips (edgeline).

• Treatments mentioned that had not been included on the mailout survey are:� butterfly reflectors within the W-beam rail of a guardrail (being used in Odessa and

Corpus Christi),� signal on high center to improve visibility for vehicles on a crest vertical curve (Austin

District), and� reflective red/white alternating material on stop sign post (Austin District).

3-1

CHAPTER 3

LITERATURE REVIEW

Just prior to the start of this Texas Department of Transportation project, the TexasTransportation Institute completed a project for the National Cooperative Highway ResearchProgram (NCHRP) that developed an Accident Mitigation Guide for Congested Rural Two-LaneHighways (NCHRP Report 440) (1). A comprehensive literature search was conducted as part ofthe NCHRP project. Therefore, this project focused on research conducted during or after theNCHRP project was active and on literature that addressed low-volume (ADT � 2000), ruraltwo-lane highways. Appendices A to I contain information on treatments for crashes on ruraltwo-lane highways from the NCHRP Report 440 report along with information identified sincethe national study.

OTHER VALUABLE REFERENCES

During the literature review process, several documents were identified that dealt with issuesother than crash treatments that may be of value when evaluating the conditions at a location. Brief summaries of these documents follow with an overview of the NCHRP research findingsand document. Also included is an overview of the Interactive Highway Safety Design Model(IHSDM) that the Federal Highway Administration (FHWA) is developing.

Interactive Highway Safety Design Model

The Federal Highway Administration has developed a software program called the InteractiveHighway Safety Design Model (IHSDM) in cooperation with state departments of transportationand several vendors of computer-aided design and other software (2). It should enable highwaydesigners to evaluate the safety of specific geometric designs for rural two-lane highways. Theestimated date of completion for the model is 2002. The IHSDM will consist of the followingseven evaluation modules:

• The policy review module allows designers to compare a proposed horizontal alignment withstate and local design standards. If a curve’s radius or superelevation deviates fromrecommended standards, relevant policy information is provided as well as a form that allowsdesigners to explain why an exception may be merited.

• The crash data module provides users information on how proposed design features willincrease or decrease the number and severity of crashes on a given stretch of road. Bymanipulating such factors as shoulder width, vehicle-per-day usage, and percentage ofcommercial vehicles on the road, designers can more accurately predict a road’s safetyrecord.

• The design consistency module will predict how a roadway alignment will affect operatingspeed.

3-2

• A driver/vehicle module will estimate vehicles’ lateral acceleration, friction demand, androlling potential.

• An intersection diagnostic review module will evaluate intersection design alternatives andidentify possible countermeasures when geometric elements compromise driver safety.

• A roadside safety module will perform cost/benefit analyses of roadside design alternatives.• A traffic analysis module will estimate how roads will perform under current and projected

traffic flows using traffic simulation models.

The current software evaluates only two-lane highway designs. FHWA hopes to develop asecond version of the program for multilane roads by 2006.

Causal Factors for Accidents on Southeastern Low-Volume Rural Roads (3)

Crashes from Kentucky and North Carolina from 1993 to 1995 were used to identify therelationship between driver, roadway, and environmental factors involved in crashes on low-volume roads. The analysis used the quasi-induced exposure technique, which identified driverand vehicle groups that are most at crash risk on rural, low-volume roads. Specific findings andconclusions include the following:

• In general, the crash trends observed for low-volume roads in Kentucky and North Carolinaare similar to trends observed on other roads.

• Young drivers, under the age of 25, show higher crash ratios for single-vehicle crashes thanany other group of drivers and are more likely to be involved in a single-vehicle crash onlow-volume roads than any age group of drivers.

• The general trend of age differences was noted for two-vehicle crashes on low-volume roads. Therefore, middle-age drivers are safer than younger drivers, who in turn are safer than olderdrivers.

• For single-vehicle crashes, the differences among age groups are larger for crashes occurringat night and on roadways with higher speeds, narrowest lanes, both narrowest and widestshoulder widths, sharpest curves, and low-volume roads. In general, younger drivers werethe least safe under all of these conditions.

• Shoulder width and roadway curvature showed that drivers have lower crash rates on roadswith the worst conditions—no shoulder or sharpest curves—than on less dangeroussegments. These data indicated that drivers increase their attention and lower theirspeeds—drive more safely—in adverse traffic environments, but they may drive lesscarefully in safer environments.

• Older drivers are less safe than younger and middle-aged drivers on roads with sharp curves,being involved in both single- and two-vehicle crashes.

• For two-vehicle crashes, the age differences are present and stronger than the roadway speedlimit, lane and shoulder width, and curvature. The data analyzed show that these factors didnot significantly affect the occurrence of two-vehicle crashes on low-volume roads.

• Female drivers are safer than male drivers. Moreover, younger female drivers are safer thanyounger male drivers, but older male drivers are safer than older female drivers. Femaledrivers from North Carolina have lower crash ratios than their Kentucky counterparts.

• Newer vehicles are more likely to be involved in single-vehicle crashes and are more likely todo so when driven by younger drivers.

3-3

• Older drivers are more likely to benefit from the increased safety levels of newer vehicles, atrend holding for both single- and two-vehicle crashes.

• Larger vehicles are more likely to hit other vehicles on the typically narrow low-volumeroads, but smaller vehicles are more likely to be involved in single-vehicle crashes.

On the basis of these findings, a series of potential countermeasures is proposed that couldimprove the traffic safety of low-volume roads.

• Most of the findings indicate that younger drivers have higher crash ratios for single vehiclesin all traditional geometric features of such roads: sharp curves, narrow lanes, no shoulders,and high speed limits. Driver education and graduated licensing appear to be the reasonablecountermeasures for improving the safety of these drivers.

• Most of the countermeasures should focus on addressing the issue of single-vehicle crashes,because more than one-half of the crashes on low-volume roads are such crashes. Short-term solutions should focus on increased driver education as well as lowering the speed limiton certain roadway segments, because all age groups of drivers have their higher crash ratioson such roads. Long-term solutions include geometric improvements dealing with increasinglane and shoulder widths and eliminating sharp curves—all geometric features contributing tothe occurrence of single-vehicle crashes.

• A number of socioeconomic characteristics may explain part of the crash rates on low-volume roads. Obviously, older vehicles are less safe than newer vehicles, and the age of thevehicle is closely tied to a variety of social factors. The data here show that the age of thevehicle is inversely proportional to the single-vehicle crash involvement and proportional totwo-vehicle crash involvement. Although newer vehicles are safer and have added safetyfeatures, compared with older vehicles, they also could be viewed as a means to reduce thesafety margins set by the drivers. This is particularly true for the younger drivers in single-vehicle crashes. These facts could be presented within a driver education program where thepotential perils of new vehicles could be demonstrated. Older vehicles present the other endof the problem where antiquated vehicles still drive on secondary, low-volume roads. Vehicle inspection programs may be an added countermeasure where vehicles with safety-related deficiencies could be identified.

National Cooperative Highway Research Program Project

Accident Mitigation Guide for Congested Rural Two-Lane Highways (NCHRP Report 440)

While the NCHRP project had several tasks and objectives, its primary purpose was the creationof an Accident Mitigation Guide for Congested Rural Two-Lane Highways (1). The AccidentMitigation Guide was to be developed to provide assistance to the transportation practitioner inidentifying and designing projects to improve safety on congested rural and exurban two- andthree-lane highways. A synopsis of the material in the Accident Mitigation Guide is provided inTable 3-1.

3-4

Chapters 3 to 6 of the Accident Mitigation Guide contain the bulk of information. They discusscountermeasures that are appropriate for congested rural and exurban two- and three-lanehighways. Each countermeasure section starts with an overview, such as a brief discussion onthe need for adequate recovery distance along a roadway. This discussion is then followed bythree subsections: Accident Experience, Countermeasures, and Effectiveness of Countermeasure. Accident Experience contains available information on the types of accidents and/or thefrequency of accidents for the situation. Appropriate countermeasures for use are discussed next. This discussion presents general information about countermeasures, techniques that are used,and examples. The final subsection discusses the known effectiveness of the countermeasures. In some cases, the effectiveness of a countermeasure is well known, such as the addition ofshoulders. In other cases, the effectiveness is suspected or not known.

To select the roadway projects that would be investigated as part of the research, a preliminarylist of potential improvements were developed and included with a mailout survey. Respondentsindicated which improvements have been implemented and where. The panel for the researchproject provided additional guidance on which types of improvements should be targeted duringthe selection process. For example, previous research has demonstrated the benefits of passinglanes and turn lanes. Therefore, the efforts were focused on other types of treatments, such asrumble strips and traveler information. The roadway projects included in the Accident MitigationGuide illustrate the types of improvements that have actually been implemented by state andlocal highway agencies that are less costly than widening the roadway to four lanes. Table 3-1(see section on Chapter 7: Examples of Safety Improvements in Table 3-1) includes a list of theprojects.

Role of Congestion in Accident Experience

An investigation to determine the role of congestion in traffic accidents on two-lane highwayswas undertaken in the research (4). The investigation used traffic volume and accident data forselected two-lane highway sites in five states. Accident frequencies, accident rates, accidentseverity distributions, and accident type distributions were determined for the sites in each stateas a function of traffic operational level of service (LOS). The conclusions of this evaluationwere:

� There is no clearly defined relationship between accident rate per million vehicle-kilometersand level of service. Different trends were found in different states, and no definitiveconclusions could be reached.

� The proportion of fatal and injury accidents increases as congestion increases under daytimeconditions. The proportion of fatal and injury accidents is lowest at LOS A (45.0 percent), ishigher for LOS B through E (53.6 percent), and is highest for LOS F (69.1 percent).

� The proportion of multiple-vehicle accidents increases and the proportion of single-vehicleaccidents falls as congestion increases for LOS B through E under daytime conditions. Thetrend of increasing multiple-vehicle accidents with increasing congestion is primarily due toincreases in the proportions of rear-end and sideswipe collisions as congestion increases.

3-5

Table 3-1. Synopsis of Material in the Accident Mitigation Guide for Congested Rural Two-Lane Highways.

Chapter 1: Introduction. This chapter discusses theneed for the Accident Mitigation Guide along withinformation on accident characteristics and the role ofcongestion on rural two-lane highways.Chapter 2: Accident Mitigation Process. Theaccident mitigation process was divided into six steps:identify sites with potential safety problems;characterize accident experience; characterize fieldconditions; identify contributing factors andappropriate countermeasures; assess countermeasuresand select most appropriate; and implement countermeasure and evaluate effectiveness. Chapter 3: Roadway Countermeasures. Theroadway chapter discusses the following two-lane ruralroadway cross section elements: lanes and shoulders,passing improvements, two-way left-turn laneimprovements, and bridges. Alignment is discussedwithin the following sections: horizontal alignment,vertical alignment, and combined alignment. Devicesthat can impact the operations and safety along a two-lane roadway is discussed in the following sections:traffic control devices and rumble strips. Chapter 4: Roadside Countermeasures. Thecondition of the roadside can affect accident frequencyand severity, especially when considering the highpercentage of accidents, particularly on rural two-laneroads, which involve a run-off-road vehicle. Theroadside chapter provides information on: recoverydistance, side slopes, obstacles, and utility poles.Chapter 5: Intersection Countermeasures. Thesections within the intersection chapter discusscountermeasures related to intersection configurationand geometry (such as type of intersection, severegrades, and angle of intersection), sight obstructions,turning improvements, and traffic control devices.

Chapter 6: Other Countermeasures. The previousthree chapters focus on different physical areas(roadway, roadside, or intersection). Factors otherthan the physical area of a highway also relate toaccidents and, in many cases, can provide the key toreducing accidents at a location or along a section ofhighway. This chapter describes the accidents andrelated countermeasures for these other factorsassociated with different types of accidents. Discussions occur on the following: speedenforcement, technology-based improvements, workzones, special events, public information andeducation, access management, older drivers,pedestrians, animals, and lighting.Chapter 7: Examples of Safety Improvements. This chapter contains information on 13 implementedimprovements: Rural Advanced Traveler InformationSystem; Innovative Electronic Advanced WarningSystem; Centerline Rumble Strips and Inverted ProfileThermoplastic Edgelines; Inverted Centerline RumbleStrips and Right- and Left-turn Channelization;Rumble Strips; Rumble Strips, Lane Striping, andGuardrail Installations; Open-Graded AsphaltConcrete Overlay; Flashing Advanced WarningBeacons for an All-way Stop Controlled Intersection;Cooperative Safety Program; Left-turn Channelizationand Pavement Rehabilitation; Left-turnChannelization; Climbing Lanes; and Addition ofPaved Shoulder and Left-turn Channelization toIncrease Roadway Width.Chapter 8: Suggested Readings. This chapterpresents an annotated list of material that cansupplement the discussions in Chapters 3 to 6 oncountermeasures. It is subdivided into referencematerials and research reports and/or papers.

These findings provide guidance for congested two-lane highway sites. First, although there isno clear relationship of accident rate per vehicle-kilometer to congestion level, accidentfrequencies clearly increase with increasing traffic volume. Therefore, congested sites are likelyto have more accidents than uncongested sites, and installation of accident countermeasures arelikely to have higher safety benefits. For example, a countermeasure that generally reducesaccidents by 20 percent will reduce more accidents at a congested site than at an uncongestedsite. Second, congested sites have a greater proportion of severe accidents than uncongestedsites. This increases both the seriousness of the safety problem at congested sites and thepotential benefits of safety countermeasures. Third, as congestion increases, the occurrence ofmultiple-vehicle accidents becomes more and more predominant. In setting priorities forimprovement, this implies that roadside improvements (which generally address single-vehiclerun-off-road accidents) are potentially desirable at any site—congested or uncongested—while

3-6

roadway improvements that address multiple-vehicle accidents become increasingly important ascongestion increases. At congested sites, priority should definitely be given to countermeasureswith the potential to reduce rear-end and sideswipe collisions such as intersection turn lanes,two-way left-turn lanes, and passing lanes.

Findings from the NCHRP Project

The Accident Mitigation Guide provides one comprehensive document that a practitioner can useto investigate several potential countermeasures for improving safety and/or operations on ruralor exurban two- and three-lane highways. The investigation of implemented countermeasuresfound the following:

• Lower-cost treatments can be highly successful in reducing accidents and/or improvingoperations along congested rural two-lane highways.

• Public participation played a significant role in the development and selection ofcountermeasures at several sites.

• Information on the selection and installation of a treatment is not always well documented. Inaddition, detailed before-and-after studies of the effectiveness of a treatment are also sparse. There were some cases, however, where the documentation of a treatment was comprehensive.

• Several of the potential treatments were part of other, larger roadway improvement projects;therefore, it was difficult to isolate the effects of the lower-cost improvement from the effectsof the other treatments.

• Several of the treatments were viewed as temporary measures to improve safety and/oroperations until the funds could be allocated to widen the roadway to four lanes.

Roadway Safety Guide (5)

The Federal Highway Administration developed a guide which was: “...designed to provide local elected officials and other community leaders with basicinformation on improving roadway safety in their communities. Written for nonengineers, it isdesigned to be a hands-on, user-friendly document, providing community leaders with:• strategies they can use right away to begin making roads safer;• basic information to improve roadway safety in cooperation with state and local

transportation departments, highway engineers, highway safety officials, SafeCommunities groups, and other safety programs; and

• clear descriptions of key funding and decision-making processes that affect roadwaysafety.

The Guide is available on the Roadway Safety Foundation website, www.roadway.org, withupdates to assist users in their ability to respond to emerging roadway safety problems.”

The report, Roadway Safety Guide for Local Decision Makers and Community Leaders, includesthe following chapters:1. Getting Started: How to Identify Roadway Safety Problems2. Choosing Countermeasures: Best Practices3. Getting It Done4. Getting Help

www.roadway.org

3-7

Low-Cost Methods for Improving Traffic Operations on Two-Lane Roads: InformationalGuide (6)

This report is an informational guide for highway agencies on the use of low-cost improvementsto alleviate operational problems on two-lane highways. The guide addresses both passing andturning improvements that can be constructed for a lower cost than construction of a continuousfour-lane highway. The passing improvements presented include passing lanes, climbing lanes,short four-lane sections, turnouts, shoulder driving, and shoulder-use sections. The turningimprovements included are intersection turn lanes, shoulder bypass lanes, and two-way left-turnlanes.

Technology in Rural Transportation “Simple Solutions” (7)

This report contains information on simple solutions identified during the rural outreach project. The goal was to identify and describe proven, cost-effective, “low-tech” solutions for ruraltransportation-related problems or needs. Research and interviews with local level transportationprofessionals were used to identify examples of technology applications. More than 50 “simplesolutions” were identified and a subset of these solutions was selected for further investigation. Details gathered included descriptions of the benefits of the technology, the expectedimplementation process, the potential issues associated with each technology, and eachtechnology’s role in a larger scale, fully integrated rural intelligent transportation system.

Prediction of the Expected Safety Performance of Rural Two-Lane Highways (8)

The report presents an algorithm for predicting the safety performance of a rural two-lanehighway. The accident prediction algorithm consists of base models and accident modificationfactors for both roadway segments and at-grade intersections on rural two-lane highways. Thebase models provide an estimate of the safety performance of a roadway or intersection for a setof assumed nominal or base conditions. The accident modification factors adjust the base modelpredictions to account for the effects on safety for roadway segments of lane width, shoulderwidth, shoulder type, horizontal curves, grades, driveway density, two-way left-turn lanes,passing lanes, roadside design and the effects on safety for at-grade intersections of skew angle,traffic control, exclusive left- and right-turn lanes, sight distance, and driveways. The accidentprediction algorithm is intended for application by highway agencies to estimate the safetyperformance of an existing or proposed roadway. The algorithm can be used to compare theanticipated safety performance of two or more geometric alternatives for a proposed highwayimprovement. The accident prediction algorithm includes a calibration procedure that can beused to adapt the predicted results to the safety conditions encountered by any particular highwayagency on rural two-lane highways. The algorithm also includes an Empirical Bayes procedurethat can be applied to utilize the safety predictions provided by the algorithm together with actualsite-specific accident history data.

3-8

Safety Improvements for Low-Volume Rural Roads (9)

The justification of safety improvements for low-volume rural roads has been difficult. Roadblocks of a primarily economic nature have prevented the improvement of many featuresassociated with this type of road, features which have been known to have adverse safetyimplications for many years. In this report, traditional methods of developing a safety index forthese roads have been explored and found unsuitable. These methods include the correlation ofcrash rates with specific roadway features and the location where atypical number of accidentsoccur. Neither of these approaches in general are of value on low-volume (ADT �1000) ruralroads. The combination of two relatively new concepts for safety improvements is recommendedas a result of this study. They are “process-based improvements” and “low-cost safetyimprovements.” For example, one “process” is to eliminate all hazardous concrete culvertheadwalls in a district. The “low-cost” aspect relates to either breaking the headwall off atground level or building up the soil of the roadside to the level of the headwall top surface.

4-1

CHAPTER 4

VEHICLE CRASHES ON ON-SYSTEM, LOW-VOLUME, RURAL TWO-LANE HIGHWAYS IN TEXAS

A study on vehicle crashes using Texas DPS crash and TxDOT roadway inventory databases wasconducted. The study was performed to partially address the following questions pertaining toon-system, low-volume, rural roads:

(1) How often do crashes occur? (2) Where do crashes occur? (3) What types of crashes occur more often?

The chapter will summarize key results from the vehicle crash study.

Addressing the first two questions is as important as addressing the third question for severalreasons. Because of the lack of resources and a vast highway system that needs to be maintained,operated, and improved, highway engineers continue to have to juggle available resources tomake incremental safety improvements. This often requires them to make difficult decisions onthe trade-off between cost and safety and other operational objectives. Under this premise,retrofitting the entire low-volume system at once with certain potential “low-cost” improvementsmay still be infeasible to do. Thus, knowing where to improve and how to prioritize andschedule the improvement is equally important in addressing the safety problems of low-volumeroads (10).

The occurrences of vehicle crashes are quite random and sporadic across the road network (11,12). Previous experience suggests that although it is almost impossible to predict when andwhere on the network a vehicle crash will occur, it is, however, quite predictable as to how manycrashes will occur on the entire network in a large area for a relatively long period of time (e.g.,one to three years). Borrowing from this experience, in this study, vehicle crashes wereexamined at three levels of aggregations: state level, district level, and county level.

The purpose of the state-level analysis was to understand how low-volume, rural two-lanehighways as a type of roadway differ from other types of two-lane roadways in terms of theirvehicle crash rates and crash characteristics. Specifically, in this analysis, crashes on six types oftwo-lane roads were examined. These roads were categorized based on their annual average dailytraffic (ADT) volumes and area type (i.e., rural versus urban). They are:

• Rural two-lane with ADT less than or equal to 2000 vehicles per day,• Rural two-lane with ADT between 2000 and 6000 vehicles per day,• Rural two-lane with ADT greater than 6000 vehicles per day,• Urban two-lane with ADT less than or equal to 2000 vehicles per day,• Urban two-lane with ADT between 2000 and 6000 vehicles per day, and• Urban two-lane with ADT greater than 6000 vehicles per day.

4-2

The district-level analysis was to shed light on the potential time-trend and spatial patterns ofvehicle crashes on the rural low-volume roads. In addition, the statistical characteristics ofvehicle crashes were compared between districts that have high crash rates and those that havelow crash rates. This later analysis provided some insights on which types of crashes occurredrelatively more often than others and what contributing factors potentially made some districtshave higher crash rates than other districts.

County-level analysis was intended to provide more details on the spatial distribution of crashes. It also identified counties that the project team could visit within other tasks. For example, basedon county crash rates, the project team selected counties with very high rates and those with lowrates but similar number of centerline miles for further investigation. Chapter 5 presents thefindings from the site-level crash evaluation.

Before presenting the results of the crash data analysis, it is worth noting that there are concernsover the quality of non-injury and property-damage-only (PDO) crash data, especially their highnon-reporting rates. Because of these concerns, the analysis in this study was conducted withKAB crashes only. Also, for those not familiar with TxDOT, the Department is currentlystructured into 25 geographic districts that are responsible for highway development. The state’s254 counties are divided among the districts. District offices divide their work into area officesand areas into maintenance offices. Texas’ variety of climates and soil conditions places differingdemands on highways, so design and maintenance, right-of-way acquisition, constructionoversight, and transportation planning are primarily accomplished locally.

STATE-LEVEL ANALYSIS

The two-lane roadways in Texas were divided into rural or urban and then into three ADT groups(less than or equal to 2000 ADT, between 2000 and 6000 ADT, and more than 6000 ADT).While working with the crash database, it was determined that crashes at an intersection are onlyassigned to the higher class or higher volume road within the state’s database. Therefore, thenumber of crashes along a long stretch of rural two-lane highway could be undercounted becausethe crashes at the intersection may not be counted. For this project, crashes were counted once ifboth roads belong to the same ADT group and counted twice if the intersecting roads belong todifferent ADT groups (once within each group). The KAB crash frequencies, million vehiclemiles traveled, KAB crash rates, and centerline miles for the six ADT/area type groups arepresented in Table 4-1 for three years from 1997 to 1999. Observations that can be made fromthe table follow:

• While many more KAB crashes occur on rural two-lane roadways, the crash rates (measuredin KAB crashes per million vehicle miles traveled, KAB/MVMT) are higher for the urbangroups. The urban groups have a much lower number of centerline miles. For 1999, of the44,606 KAB crashes in Texas, 31 percent occurred on two-lane highways (13,909) withapproximately 75 percent of those crashes occurring in rural areas. The remaining KABcrashes (30,697) occurred on roads with more than two lanes.

• Each two-lane ADT group in both rural and urban areas in 1999 had KAB crash rates(between 36.3 and 45.7 KAB/100 MVMT for rural and 53.0 and 104.1 KAB/100 MVMT forurban) that were greater than the crash rate for all on-system roadways (31.5 KAB/100

4-3

MVMT). An interpretation of the data is to note that a vehicle traveling on two-laneroadways, whether in an urban or rural environment, has a greater likelihood of beinginvolved in a KAB crash per VMT than traveling on a multilane roadway.

• In terms of centerline miles, rural low-volume roads constitute over 79 percent of the on-system two-lane roads.

Table 4-1. Texas On-System, Two-Lane Highways.

ADT Group KAB Crashes MVMT KAB/100MVMT

CenterlineMiles

1999

R

ural

ADT��

2000ADT = 2 to 6000

ADT > 6000

482449022356

10,56112,9676496

45.737.836.3

45,67410,2682031

U

rban

ADT��

2000ADT = 2 to 6000

ADT > 6000

20510462697

19714205,086

104.173.753.0

4469761114

All Two-Lane (w/Crashes Double Count)All Two-Lane (No Crashes Double Count)

All On-System

16,03013,90944,606

36,727

141,450

43.6

31.5

60,509

73,772

1998

R

ural

ADT��

2000ADT = 2 to 6000

ADT > 6000

482246752202

10,58712,8386015

45.536.436.6

45,86510,2091906

U

rban

ADT��

2000ADT = 2 to 6000

ADT > 6000

19510442872

21114655159

92.471.355.7

4789981112


All On-System

15,81013,77744,355

36,275

138,927

43.6

31.9

60,568

73,724

1997

R

ural

ADT��

2000ADT = 2 to 6000

ADT > 6000

497646222210

10,74412,1926008

46.337.936.8

46,62997111917

U

rban

ADT��

2,000ADT = 2 to 6000

ADT > 6000

20212103069

22716155557

89.074.955.2

52411171218


All On-System

16,17314,11145,050

36,343

131,312

44.5

34.3

61,116

72,792

4-4

• Within each area type, higher volume roads tend to have lower KAB crash rates due,presumably, to better roadway design. As will be discussed later, higher volume roads havelower percentages of KAB crashes occurring on curves.

• Urban two-lane roads have significantly higher KAB crash rates than rural two-lanes. As willbe discussed later, this is due, most likely, to a higher number of intersection andintersection-related crashes on higher volume roads.

Appendix J contains the statistical characteristics of vehicle crashes for the three ADT groupswithin the rural and urban environment. These statistics are based on three years of crash recordsfrom 1997 to 1999. Frequencies and distributions of KAB crashes are shown by injury severity,whether they are intersection-related, roadway alignment, horizontal curvature, weatherconditions, lighting conditions, pavement wetness conditions, month-of-year, day-of-week, time-of-day, manner of collision, first harmful event, and object struck. Table 4-2 lists observationsfrom these distributions (with focus on the row percent, i.e., the third value in each group asshown in Appendix J):

Table 4-2. Characteristics of Crashes within Development/ADT Groups.

Category Observation

Injury Severity For the KAB crashes, within each area type, higher volume roads tend to have lowerpercentages of fatal crashes. Rural two-lane roads have significantly higher percentages offatal crashes than the urban two-lanes.

Intersection,Intersection-Related, andDriveway-RelatedCrashes

Urban two-lane roads have considerably higher percentages of intersection, driveway-related,or intersection-related crashes than the rural two-lanes. For example, urban

�2000 ADT had

62 percent while rural �

2000 ADT only had 33 percent of intersection, driveway- orintersection-related crashes. High ADT Groups have higher percentages of intersection,intersection-related, and driveway-related crashes than the 2K Group (e.g., the rural > 6000ADT had 55 percent intersection, driveway- or intersection-related crashes).

Alignment Most crashes occurred on straight, level sections (66 to 95 percent of the KAB crashes). Thepercentages of KAB crashes that occurred on curved, level road sections for each of the ADTgroups are: 32, 18, and 9 percent for rural roads and 15, 11, and 5 percent for urban roads. This suggests that the presence and/or the design of horizontal curves is a major roadwayfactor associated with low-volume roads having significantly higher KAB crash percentagesas compared to the higher volume roads. In addition, it suggests that horizontal curves are amajor factor that contributes to the higher frequencies of curve-related crashes for rural roadsthan for the urban roads.

HorizontalCurvature

A larger percentage of KAB crashes occurred on tight horizontal curves (defined as beinggreater than or equal to 4 degrees) than on larger radius curves. The percent of KAB for eachADT group for crashes on curves with a degree of curvature of 4 or more were 21, 10, and 5percent for the rural ADT groups and 11, 9, and 4 percent for the urban ADT groups. Thisfurther indicates that the existence of sharp curves in rural low-volume roads is a major factorresponsible for their higher KAB crash rates. Previous research has also found that horizontalcurves experience a higher crash rate than tangents on rural two-lane highways (13).

Weather For rural roads, higher ADT groups had a slightly higher percentage of crashes that occurredon rainy days (9.5 and 9.4 percent versus 7.2 percent).

4-5

Table 4-2. Characteristics of Crashes within Development/ADT Groups (continued).


LightingConditions

Considerably higher percentages of the KAB crashes occurred on dark, not lighted, roads forrural and for low-volume roads. For roads in rural areas with less than 2000 ADT, 37 percentof the KAB crashes occurred during dark, not lighted, conditions while only 27 percent of theKAB crashes in the urban low-volume area occurred under similar lighting conditions.

SurfaceConditions

About 14 to 16 percent of all KAB crashes regardless of area type or ADT occurred underwet/muddy/snowy conditions.

Month-of-Year Crashes occurred quite uniformly throughout the year with May, July, and October havingslightly higher percentages of crashes.

Day-of-Week For rural roads regardless of the ADT groups, more crashes occurred on Friday, Saturday, andSunday, with Saturday having the highest percentage (about 19 percent). Urban roads are,however, different. Their highest percentage is on Friday, lowest generally on Sunday, anduniform for the rest of the days.

Time-of-Day The higher percentages of KAB crashes occurred between 3 pm and 7 pm for alldevelopment/ADT groups.

Manner ofCollision/VehicleMovement

Low-volume roads have considerably higher percentages of single-vehicle crashes than high-volume roads, and rural two-lane roads have significantly higher percentages of single-vehiclecrashes than urban two-lane roads. On rural low-volume two-lane roads, 68 percent ofcrashes involve a single vehicle while only 40 percent of crashes involve a single vehicle onurban low-volume two-lane roads. At higher ADTs, the percentage for single vehicle drops to31 percent for rural and 19 percent for urban (two-lane roads with ADT over 6000).

First HarmfulEvent

For the rural 2K Group, about 61 percent of the crashes are either overturned or fixed-objectcrashes and the percentages decrease as ADT increase (42 percent for 2-6000 group and 26percent for 6000+ group). These percentages are considerably higher than the urban roads intheir respective ADT categories (which are 37, 26, and 15 percent, respectively). For urbanroadways and higher volume rural roadways (>2000), the majority of the crashes involvedstriking another moving vehicle. Only 31 percent of the crashes on rural 2K roads involvedstriking another moving vehicle.

Object Struck Rural roads and low-volume urban roads have much higher percentages of tree/shrub, fence,and culvert/headwall crashes. Low-volume urban roads also have a high percentage of utility-pole crashes. For low-volume rural two-lane roads the type of object struck is: no codeapplicable (50 percent), fence (13.5 percent), tree/shrub (9.7 percent), culvert/headwall (5.0percent), highway sign (3.7 percent), embankment (2.5 percent), ditch (2.5 percent), otherfixed object (2.3 percent), and utility pole (2.1 percent). All other objects had percentagesless than 2.

Other Factor Only 29 percent of the accidents had an “other factor” code used. Codes used were attentiondiverted (4.1 percent), swerves due to animal (4 percent), moving vehicle entering driveway(3.1 percent), moving vehicle pass on left (2.1 percent), and highway under construction (2.1percent).

4-6

DISTRICT-LEVEL ANALYSIS

Figure 4-1 shows KAB crash rates on on-system, low-volume (less than or equal to 2000 ADT),rural two-lane highways for each TxDOT district from 1992 to 1999. Based on these rates, thedistricts were grouped into three “rate groups.”

• High-Rate Group: Atlanta, Austin, Bryan, Dallas, Ft. Worth, Houston, Lufkin, and Tyler;• Mid-Rate Group: Beaumont, Brownwood, Corpus Christi, Paris, Pharr, San Antonio, Waco,

Wichita Falls, and Yoakum; and• Low-Rate Group: Abilene, Amarillo, Childress, El Paso, Laredo, Lubbock, Odessa, and San

Angelo.

Figure 4-2 shows the location of the rate groups in the state. Several interesting observationscould be made with regards to the crash-rate time series shown in Figure 4-1:

• The Lufkin District has one of the highest KAB rates in the previous four years. The DallasDistrict had a large increase in its KAB crash rate between 1998 and 1999.

• The eight districts in the High-Rate Group have higher than average rates consistentlythroughout each of the nine years while the eight low-rate districts consistently have below-average rates throughout the same period.

• Pharr District shows a significant drop in crash rate in and after 1996 while Wichita Fallsexperienced a jump in crash rate in and after 1996.

• The overall KAB crash rate was about 0.4 crashes per million vehicle miles (MVM). TheHigh-Rate Group has an average rate of about 2.5 times higher than that of the Low-RateGroup.

0.000

0.100

0.200

0.300

0.400

0.500

0.600

0.700

0.800

0.900

1990 1992 1994 1996 1998 2000

Year

KA

B C

rash

Rat

e (c

rash

es/M

VM

)

FORT WORTH

TYLER

LUFKIN

HOUSTON

AUSTIN

BRYAN

DALLAS

ATLANTA

A. High-Rate Group.Figure 4-1. KAB Crash Rates (Per Million Vehicle Miles Traveled) by District, 1992 to

1999, On-System, Low-Volume, Rural Two-Lane Highways.

4-7

0.000

0.100

0.200

0.300

0.400

0.500

0.600

0.700

0.800

0.900

1990 1992 1994 1996 1998 2000

Year

KA

B C

rash

Rat

e (c

rash

es/M

VM

)

AMARILLO

LUBBOCK

ODESSA

SAN ANGELO

ABILENE

LAREDO

EL PASO

CHILDRESS

0.000

0.100

0.200

0.300

0.400

0.500

0.600

0.700

0.800

0.900

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Year

KA

B C

rash

Rat

e (

cras

hes

/MV

M)

PARIS

WICHITA FALLS

WACO

YOAKUM

SAN ANTONIO

CORPUS CHRISTI

BEAUMONT

PHARR

BROWNWOOD

B. Mid-Rate Group.

C. Low-Rate Group.

Figure 4-1. KAB Crash Rates (Per Million Vehicle Miles Traveled) by District, 1992 to1999, On-System, Low-Volume, Rural Two-Lane Highways (continued).

4-8

KAB Rate by District

Low - RateMid - RateHigh - Rate

Figure 4-2. Location of Crash Rate Groups in Texas.

Appendix K contains the statistical characteristics of vehicle crashes for the three district groups. These statistics are based on three years of crash records from 1997 to 1999, two-lane, rural, andADT less than or equal to 2000. The frequencies and distributions of KAB crashes are presentedin similar groups as used for the state-level data. Table 4-3 lists observations from thesedistributions (with focus on the row percent, i.e., the third value in each group).

4-9

Table 4-3. Characteristics of Crashes within KAB Rate Groups.


Injury Severity Low-Rate Group has a slightly higher percentage of fatal crashes (10.6 percent) than theHigh-Rate Group (8.7 percent) and the Mid-Rate Group (9.1 percent).

Intersection,Intersection-Related, andDriveway-RelatedCrashes

More of the crashes in the Low-Rate Districts were not related to an intersection (73 percent)than in the Mid-Rate Districts (65 percent) and the High-Rate Districts (65 percent). TheHigh-Rate Districts have more driveway-related crashes (10 percent) than the Mid-RateDistricts (8 percent) and the Low-Rate Districts (5 percent).

Alignment Approximately 37 percent of the crashes in the High-Rate Group occurred on level, horizontalcurves. This is much higher than the Low-Rate Group (24 percent) and Mid-Rate Group (31percent). This indicates that horizontal curves are a major contributing factor to crashes onsites within the High-Rate Group.

HorizontalCurvature

The majority of crashes on curves are occurring on tight curves (greater than or equal to 4degrees). For the High-Rate Districts, 25 percent occurred on 4 degree or more curves, 16percent were on curves with less than 4 degrees, and the remainder were on no curve orunknown. The Mid-Rate Districts also had a similar pattern with 20 percent occurring on 4degree or more curves, 14 percent on curves with less than 4 degrees, and the remainder on nocurve or unknown. The Low-Rate Districts had similar percentages of curves for more than 4degrees (13 percent) and less than 4 degrees (12 percent). These findings further suggest thatthe existence of sharp curves is a significant contributing factor on two-lane rural highways.

Weather For all groups, about 89 percent of the crashes occurred on clear or cloudy days.

LightingConditions

There is a very small difference between the different groups in terms of the percentage ofcrashes that occurred under dark/dawn/dusk conditions. Overall, about 43 percent of thecrashes occurred under these conditions. With such a high percentage of crashes occurring inthese conditions, low-cost improvements to reduce nighttime crashes should be considered.

SurfaceConditions

The High-Rate Group had a slightly higher percentage of crashes occurring under wetpavement than the Low-Rate Group (14.3 versus 10.1 percent). The Low-Rate Group had ahigher percentage of crashes occurring on snowy conditions (3.4 percent) than the High-RateGroup (0.4 percent) or the Mid-Rate Group (0.9 percent).

Month-of-Year More crashes occurred in May, July, and October for all groups with the lowest percentage ofcrashes occurring in February.

Day-of-Week More crashes occurred on Friday, Saturday, and Sunday, with Saturday having the highestpercentage (over 18 percent for each group).

Time-of-Day Similar observations can be made as in the Lighting Conditions.

Manner ofCollision/VehicleMovement

All three rate groups have similar distributions.

4-10

Table 4-3. Characteristics of Crashes within KAB Rate Groups (continued).


First HarmfulEvent

The High- and Mid-Rate Groups had a higher percentage of fixed object crashes (35 and 33percent) than the Low-Rate Group (25 percent). The Low-Rate Group had a higherpercentage of overturned crashes (39 percent) than the other groups (26 percent for High-Rategroup and 27 percent for Mid-Rate Group). With such a high percent of overturned/fixedobject crashes (over 60 percent for each group), improvements to keep the vehicles on theroad and maintain vehicle stability both on-road and off-road are critical. The data also showthat approximately 4 to 6 percent of the crashes involved an animal as the first harmful event.

Object Struck The top three types of objects that vehicles struck were tree/shrub, fence, andculvert/headwall. For the High-Rate Group, the percentages were 13.3, 12.6, and 5.8 percent,respectively, while for the Low-Rate Group, these percentages were 2.6, 13.7, and 4.0percent, respectively. This finding demonstrates that tree/shrub are important characteristicsof the High-Rate Group.

Other Factor The three district groups had similar distributions for the Other Factors category. They reflectthe low-volume nature of the roadways. Most of the crashes had no code applicable (70 to 72percent). Codes that were selected included attention diverted (3.6 to 4.9 percent), swervingto miss an animal (4 percent), and moving vehicle entering driveway (2.1 to 3.5 percent).

COUNTY-LEVEL ANALYSIS

The county-level analysis was intended to provide more details on the spatial distribution ofcrashes. It was also intended to provide an indication on which counties the project team maywant to visit or select when conducting more detailed engineering analyses. Seven years of crashrecords and road inventory data from 1992 to 1999 were examined for each county.

Figure 4-3 shows the number of centerline miles by county (averaged over the seven-yearperiod). The 37 counties with the highest number of centerline miles are listed in Table 4-4. Figure 4-4 presents KAB crash rates by county (in crashes per 100 MVM), and the 45 countieswith the highest rates are listed in Table 4-5. The darker the shading in Figure 4-3, the higher thecrash rate. The figure illustrates that higher KAB crash rates are present in the eastern portion ofthe state.

4-11

Centerline Miles

0-112113-225226-337338-450

Figure 4-3. Centerline Miles by TxDOT County for On-System, Low-Volume, Rural Two-Lane Highways (average for 1992 to 1999).

4-12

KAB Crashes/100 MVM

0-40

41-80

81-120

121-160

Figure 4-4. KAB Crashes/100 MVM by TxDOT County for On-System,Low-Volume, Rural Two-Lane Highways (average for 1992 to 1999).

4-13

Table 4-4. Top 37 Counties with the Most Centerline Miles (8-year Averages)(On-System, Low-Volume, Rural Two-Lane Highways, 1992-1999).

County Name County Number District Name Centerline Miles

Pecos 186 Odessa 449

Fannin 75 Paris 378

Cherokee 37 Tyler 366

Jones 128 Abilene 343

Navarro 175 Dallas 341

Rusk 201 Tyler 336

Lamar 139 Paris 328

Hill 110 Waco 327

Red River 194 Paris 317

Cass 34 Atlanta 317

Hunt 117 Paris 317

Houston 114 Lufkin 304

Gonzales 90 Yoakum 303

Wood 250 Tyler 303

Lamb 140 Lubbock 302

Young 252 Wichita Falls 299

Eastland 68 Brownwood 295

Bosque 18 Waco 291

Karnes 129 Corpus Christi 290

Van Zandt 234 Tyler 288

Fayette 76 Yoakum 287

Runnels 200 San Angelo 287

Hale 96 Lubbock 286

Anderson 1 Tyler 283

Limestone 147 Waco 282

Floyd 78 Lubbock 280

Clay 39 Wichita Falls 280

Reeves 195 Odessa 278

Commanche 47 Brownwood 277

Brewster 22 El Paso 277

Coleman 42 Brownwood 276

Montague 169 Wichita Falls 275

Lynn 153 Lubbock 272

Shelby 210 Lufkin 272

Atascosa 7 San Antonio 271

Lubbock 152 Lubbock 271

Ellis 71 Dallas 271

4-14

Table 4-5. Top 45 Counties with the Highest KAB Crash Rate per MVMT.(Low-Volume, Rural Two-Lane Highways, 1992-1999).

County Name County Number District Name Centerline Miles KAB Crash Rate

Travis 227 Austin 37 1.508Harris 102 Houston 3 1.183

Somervell 213 Fort Worth 55 1.097Angelina 3 Lufkin 195 1.083Rockwall 199 Dallas 53 1.070

Gregg 93 Tyler 50 1.023Montgomery 170 Houston 86 0.919Washington 239 Bryan 167 0.889

Brazoria 20 Houston 52 0.866Smith 212 Tyler 227 0.855Bexar 15 San Antonio 66 0.846Polk 187 Lufkin 225 0.838

Galveston 85 Houston 7 0.834Shelby 210 Lufkin 272 0.832

Johnson 127 Fort Worth 136 0.831Guadalupe 95 San Antonio 169 0.822Cameron 31 Pharr 220 0.804

Lee 144 Austin 117 0.800Williamson 246 Austin 215 0.797

Camp 32 Atlanta 86 0.787Hood 112 Fort Worth 85 0.780

Harrison 103 Atlanta 253 0.776Kendall 131 San Antonio 104 0.773Orange 181 Beaumont 41 0.769Burnet 27 Austin 171 0.767

Nacogdoches 174 Lufkin 216 0.765Bastrop 11 Austin 166 0.728Upshur 230 Atlanta 210 0.724

Kerr 133 San Antonio 161 0.720Burleson 26 Bryan 157 0.718Kaufman 130 Dallas 206 0.712

Henderson 108 Tyler 225 0.707Nueces 178 Corpus Christi 142 0.692

San Jacinto 204 Lufkin 177 0.686Hays 106 Austin 56 0.684

Bandera 10 San Antonio 153 0.682Mclennan 161 Waco 238 0.682

Titus 225 Atlanta 89 0.681Walker 236 Bryan 155 0.678Rusk 201 Tyler 336 0.673

Victoria 235 Yoakum 107 0.668Bell 14 Waco 203 0.655

Waller 237 Houston 126 0.645Collin 43 Dallas 225 0.642Hunt 117 Paris 317 0.642

4-15

SUMMARY OF FINDINGS

This study found the following for the three questions asked at the beginning of the chapter:

(1) How often do crashes occur?

In 1999, there were 45.7 KAB crashes/100 MVMT on low-volume, rural two-lanehighways. For all on-system roads, the rate was 31.5 KAB/100 MVMT. For 1999, of the44,606 KAB crashes in Texas, 31 percent occurred on two-lane highways withapproximately 75 percent of those crashes occurring in rural areas. Approximately 11percent of all KAB crashes in Texas in 1999 occurred on low-volume (�2000 ADT), ruraltwo-lane highways.

(2) Where do crashes occur?

More KAB crashes occurred in eastern counties (see Figure 4-3) than western counties. The crash rates revealed that a vehicle traveling on a two-lane road, whether in an urban orrural environment, has a greater likelihood of being involved in a KAB crash per VMT thantraveling on a multi-lane highway.

(3) What types of crashes occur more often?

In general, crashes on low-volume, rural two-lane highways occur between intersections, bya single vehicle running off the road and then overturning or striking a fixed object (fence,tree/shrub, culvert). Crashes on curves (level) and in dark, non-light conditions are morecommon on low-volume, rural two-lane highways than on urban roads.

Based upon the findings from the comparison of the crashes at the state and district levels, thefollowing are key directions a district may want to pursue when considering various types of low-cost improvements:

• treatments that either decrease the number of vehicles from leaving the roadway, especiallyon tight horizontal curves or that better communicate the nature of the curve;

• improvements to reduce the number of nighttime crashes;• treatments that reduce crashes at driveways; and• improvements to minimize severity of crashes if a vehicle leaves the road.

5-1

CHAPTER 5

EVALUATION OF CRASHES AT ROADWAY LEVEL

When the accident rates by county were plotted, a definite pattern of areas with high rates versusareas with lower rates emerged (see Figure 4-2). The counties with the higher crash rates arelocated in the eastern portion of Texas. With only a few exceptions, most of the lower crashrates were found in west Texas. Known characteristics between east and west Texas that wouldcontribute to this pattern include the pine forests of east Texas versus the deserts of west Texasand the typical cross section and alignment associated with the age of the roads in the areas. Older, rural roads in east Texas are assumed to be more narrow and more curvilinear ascompared to the rural roads in west Texas. To identify whether these assumptions are valid andto identify if other roadway characteristics are associated with the different regions, a sample ofcounties was selected to investigate which regional characteristics are associated with high- andlow-crash rates.

SITE SELECTION

The two counties with the highest average KAB rates on low-volume, rural two-lane highwaysfor 1992 to 1998 were identified: Angelina and Travis. Selecting two western counties with lowKAB rates for comparison could result in a county that has a low KAB rate because it only had afew miles that met the less than 2000 ADT criteria. If so, then the difference in KAB rate couldbe because of the lack of opportunity for a crash (because of the low number of miles) rather thana true difference between the east and west regions. To control for that issue, counties that had asimilar number of miles of low-volume, rural two-lane roads to Angelina and Travis Countieswere identified. Because the number of miles that would be considered low volume will changeas cities develop and expand, the data for the most recent year available (1998) were used toidentify counties. Martin County with 185 miles of low-volume, rural two-lane roads wasmatched to Angelina County (189 miles). Travis County with 22 miles was matched to El PasoCounty (35 miles). Figure 5-1 shows the location of the four counties, and Figure 5-2 includespictures of one of the study sites within each county.

Another advantage to the Travis County and El Paso County pair is that both counties includemedium-sized, growing cities. Travis County is home to Austin, and El Paso County is home tothe city of El Paso. The growth of these cities is causing increased traffic volumes on rural two-lane roadways. Therefore, in addition to El Paso County having centerline miles similar toTravis County, they both share characteristics common to an area undergoing development. Thecity of Lufkin is the population center of Angelina County while Martin County is apredominantly rural county in the Odessa District, which has one of the lowest KAB rates.

5-2

Figure 5-1. Counties in Texas Selected for Analysis.

Approximately 20 to 30 miles of roads within each county with the highest number of crasheswere identified. The sites were initially identified by highway number and control section. Aspart of the data collection effort, the research team gathered roadway characteristics for eachcontrol section. During the trips to El Paso and Travis Counties, it was determined thatsignificant portions of two of the sites had been expanded to four lanes and/or had ADTs muchhigher than 2000. These locations typically occurred either near an intersection or interchangewith a higher functional class road or near a town. Locations with four lanes were removed fromthe study. Most of the locations with ADTs over 2000 were also removed; although one sectionin El Paso with an ADT of 4188 was retained so that a similar number of miles would beavailable between El Paso and Travis Counties. Table 5-1 lists the sections used in the analysis,along with their lengths, yearly ADTs, vehicle miles traveled, and three-year crash totals.

5-3

Martin County Angelina County

El Paso County Travis County

Figure 5-2. Samples of Study Sites in Four Texas Counties.

DATA COLLECTION

Two primary sources of data were collected for this evaluation: site characteristics data and crashrecord data.

Crash Records

TxDOT maintains the crash records for the state using information provided by the Departmentof Public Safety (DPS). A copy of the electronic crash data files is also available within theTexas Transportation Institute. These files identify the characteristics of the crashes on thesections identified within the four counties for the three-year period of 1997 to 1999. During theinitial evaluation of the crash data, it was determined that crashes at an intersection are onlyassigned to one road. If the roads are two different classes (i.e., Interstate, US, SH, or FM), thecrash is assigned to the higher-class road. If the roads are of the same class, the crash is assignedto the lower-numbered road. Since nearly all of the control sections within this study are high-numbered FM roadways, the crashes at intersections were almost always assigned to the crossingroad. Therefore, the number of crashes initially included in the analysis was undercounted.

5-4

Characteristics of the crashes assigned to the cross road for each intersection with our study siteswere obtained and used within the analysis. To allow comparison between the crashes on thesections selected for this evaluation and all crashes on low-volume, rural two-lane highways (seeChapter 4), this study examined KAB crashes only.

Table 5-1. Sections Selected for Field Investigation.Ctrl Sect Route Length

(mi)ADT MVMT Number of Crashes a KAB/

MVMTN C B A K KABANGELINA COUNTY

336-9 FM 1669 1.8 213 0.4 0 0 2 0 0 2 4.8390-4 FM 1270 6.0 210 1.4 0 0 0 2 0 2 1.51874-2 FM 2021/

FM 35213.7 687 2.8 6 2 4 4 1 9 3.2

1874-1 FM 2021 8.5 876 8.2 5 6 21 7 1 29 3.62115-1 FM 2251 5.4 1317 7.8 3 0 7 3 1 11 1.4

Totals 25.4 20.5 14 8 34 16 3 53 2.6

MARTIN COUNTY1871-2 FM 2002 6.0 106 0.7 0 0 0 0 0 0 0.01638-2 FM 829 15.8 387 6.7 0 0 0 1 0 1 0.1

494-3 (B) SH 137 12.0 1269 16.7 1 0 3 0 0 3 0.2Totals 33.8 24.1 1 0 3 1 0 4 0.2

EL PASO COUNTY2-3 (C-D) SH 20 7.9 1350 11.7 0 0 3 1 0 4 0.3

674-2 FM 76 5.6 1991 12.2 0 0 3 1 0 4 0.32-15 FM 1109 4.1 1433 6.4 0 0 0 0 0 0 0.0

2326-1 (C) FM 2529 5.0 4188 22.9 0 0 0 0 0 0 0.0Totals 22.6 53.3 0 0 6 2 0 8 0.2

TRAVIS COUNTY2210-1 RM 2322 4.6 2106 10.6 8 3 3 1 1 5 0.52718-1 RM 2769 7.3 2313 18.5 2 4 11 1 1 13 0.71378-1 RM 1431 9.8 907 9.7 0 0 4 1 0 5 0.5

Totals 21.7 38.8 10 7 18 3 2 23 0.6a N= Non-Injury C = Possible Injury B = Non-Incapacitating A = Incapacitating K = Fatal KAB = Fatal, Incapacitating, or Non-Incapacitating

Site Characteristics Data

In order to fully appreciate the characteristics of the sections chosen for evaluation, it wasnecessary to visit the sections in person and record information about basic features. Data werecollected at each site by driving the site and recording data using video or noting characteristicson a pre-developed data collection sheet (see Table 5-2 for a list of the information collected). The first round trip was a simple observation trip to become familiar with the site and take noteof any unique elements. On the second trip, a video record of the entire length of the control

5-5

section was made. While one technician was driving through the site, the second technician useda camcorder to record the view through the front windshield to obtain a driver’s point of view ofthe roadway. A third round trip was made to fill out the sections of the worksheet for roadsideenvironment, roadside development, and traffic control devices. A fourth round trip was used tocount driveways and intersections on both sides of the road, and to count vertical curves andadvisory speeds. Depending on the access density or the terrain, it was sometimes necessary todivide these tasks between two round trips. A final trip through the site provided the opportunityto stop and measure lane and shoulder widths and take pictures of both directions of the controlsection.

Table 5-2. List of Data Collected within the Site Characteristics Worksheet.

Date:Route:City/County:Control Section:

Beginning Feature:Ending Feature:Length:Direction 1: NB SB EB WB

• Roadside environment (measured for Directions 1and 2 at both 2 ft and 10 ft from pavement):• No fixed objects• Yielding objects only• Combo of yielding and isolated rigid objects• Isolated rigid objects only• Many or continuous rigid objects

• Roadside development• Trees• Farmland• Residential• Commercial• Park/School/Campus

• Number of access points along section (drivewaysand intersections)

• Lane width (by direction)• Shoulder type and width (by direction)• Total pavement width• Number of traffic signals and stop signs• Posted speed limit• Number of advisory speeds (and values)• Presence of pavement markings• Notes of any sight distance restrictions, unusual

features, and unique characteristics

DATA ANALYSIS

Crash Records

Table 5-3 contains the KAB crash data from the crash database for the study sites, grouped bycounty. The total statewide numbers are included for comparison; these data reflect 14,742 KABcrashes on rural two-lane highways with less than 2000 ADT from 1997 to 1999. The 1999statewide KAB rate for low-volume, rural two-lane highways was 0.46 crashes per millionvehicle miles traveled. The crash rates for the control sections driven in the four counties variedfrom 0.15 in El Paso County to 2.58 in Angelina County (see Figure 5-3). The selected roads inAngelina County had the most crashes of any of the counties included in the study, with 53crashes. Travis County had 23 crashes, El Paso County 8, and Martin County 4.

Most of the crashes on low-volume, rural two-lane highways in Texas occur away fromintersections. Over 73 percent are coded as being non-intersection crashes. While the sectionsselected for this study also had most of the crashes coded as non-intersection (between 49 and 65percent, excluding Martin County), they did have a greater portion coded as being at an

5-6

intersection (between 25 and 100 percent) when compared to all low-volume, rural two-lanehighways in Texas (10 percent). Figure 5-4 shows the distribution of crashes by where theyoccurred with respect to an intersection for the state of Texas and for the four counties combined. Figure 5-5 shows the distribution of accidents on the roadways within each county. There were

Table 5-3. Crash Data for 1997-1999 for Control Sections, by County and Statewide.

Num % Num % Num % Num % Num Pct Num %Overall Accidents 53 4 8 23 88 14622

Intersection 21 40 4 100 2 25 5 22 32 36% 2248 15Intersection-Related 2 4 0 0 0 0 1 4 3 3% 1346 9Driveway Access 4 8 0 0 1 13 2 9 7 8% 1269 9Non-Intersection 26 49 0 0 5 63 15 65 46 52% 9759 67

Pedestrian 0 0 0 0 0 0 0 0 0 0% 145 1Another vehicle in transport 28 53 4 100 3 38 8 35 43 49% 4548 31RR Train 0 0 0 0 0 0 0 0 0 0% 35 0Parked Car 1 2 0 0 0 0 0 0 1 1% 72 0Pedalcyclist 1 2 0 0 0 0 0 0 1 1% 71 0Animal 0 0 0 0 0 0 0 0 0 0% 637 4Fixed Object 16 30 0 0 4 50 11 48 31 35% 4739 32Other Object 0 0 0 0 0 0 0 0 0 0% 44 0Overturned 6 11 0 0 0 0 4 17 10 11% 4209 29Other non-collision 1 2 0 0 1 13 0 0 2 2% 122 1

Fatal (K) 3 6 0 0 0 0 2 9 5 6% 1344 9Incapacitating (A) 16 30 1 25 2 25 3 13 22 25% 4348 30Non-incapacitating (B) 34 64 3 75 6 75 18 78 61 69% 8930 61

1 23 43 0 0 4 50 15 65 42 48% 9992 68 2 27 51 4 100 2 25 8 35 41 47% 4393 30 3 2 4 0 0 1 13 0 0 3 3% 222 2 4 1 2 0 0 1 13 0 0 2 2% 12 0

Daylight 33 62 4 100 4 50 11 48 52 59% 8338 57Dark - No Street Lights 18 34 0 0 3 38 10 43 31 35% 5344 37Dark - Street Lights 2 4 0 0 1 13 2 9 5 6% 389 3Dawn 0 0 0 0 0 0 0 0 0 0% 269 2

Dry 46 87 3 75 8 100 22 96 79 90% 12539 86Wet 7 13 1 25 0 0 1 4 9 10% 1909 13

Clear 46 87 3 75 8 100 22 96 79 90% 13054 89Raining 7 13 1 25 0 0 1 4 9 10% 1057 7

Unknown 0 0 0 0 1 13 1 4 2 2% 601 4No curve 34 64 4 100 6 75 15 65 59 67% 8853 610.1-1.9 6 11 0 0 0 0 0 0 6 7% 695 52.0-3.9 3 6 0 0 0 0 3 13 6 7% 1384 94.0-5.9 3 6 0 0 0 0 1 4 4 5% 1195 86.0 or higher 7 13 0 0 1 13 3 13 11 13% 1894 13

Injury Severity

First Harmful Event

Intersection-Related Accidents

Light Conditions

Number of Vehicles Involved

Degree of Curve

Weather Conditions

Surface Conditions

Total StatewideAngelina TravisMartin El Paso

5-7

DA8%

IR3%

Int36%

NI53%

Control Section Accidents

DA9%

IR9%

Int15%

NI67%

Texas Accidents

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1997

-199

9 Ang

elina

1997

-199

9 M

artin

1997

-199

9 El P

aso

1997

-199

9 Tra

vis

1999

Sta

tewide

KA

B C

rash

es p

er M

VM

T

almost an equal number of intersection (21) and non-intersection (26) crashes in AngelinaCounty. Fifteen of the 23 crashes in Travis County were non-intersection crashes, as were five ofthe eight crashes in El Paso County. All four crashes in Martin County were intersection crashes. As a group, the sites selected for this study have more intersection or driveway-related accidentsthan most low-volume, rural two-lane highways in Texas.

Figure 5-3. KAB Rates for Selected Roadways.

(Int = Intersection, NI = Non-Intersection, IR = Intersection-Related, DA = Driveway Access)

Figure 5-4. Distribution of Crashes by Intersection Influence.

5-8

133%

251%

38%

48%

Western Counties

0

10

20

30

40

50

60

Angelina Martin El Paso TravisCounty

3-Y

ear

Acc

iden

t T

ota

lNon-Intersection

Driveway Access

Intersection-Related

Intersection

150%2

46%

41%

33%

Eastern Counties

Figure 5-5. Crashes for Roadways by Intersection Influence.

Along with having the majority of the crashes associated with intersections, the selectedroadways had more of their crashes involving more than one vehicle (see Figure 5-6) and the firstharmful event was, in most cases, striking another vehicle. Figure 5-7 shows the distribution offirst harmful event codes by different counties and the distribution for all roads studied and forthe state. The counties in the west show that the majority of the crashes involved two or morevehicles (67 percent). Crashes in the two eastern counties (Angelina and Travis) were splitalmost evenly between one- and two-vehicle crashes. Crashes in the two western counties(Martin and El Paso) were predominantly multi-vehicle crashes. All four crashes in MartinCounty were two-vehicle collisions at intersections. For all low-volume, rural two-lanehighways only 26 percent of the crashes involved two or more vehicles.

Figure 5-6. Crashes by Number of Involved Vehicles.

5-9

FO30%

AV53%

OT11%

Other2%

PD2%

PC2%

Angelina County

FO32%

OT28%

AV33%

Other1%

PED1%

Animal5%

AV100%

Martin County

FO49%

AV38%

Other13%

El Paso County

AV50%

FO35%

OT11%

Other2%

PD1%

PC1%

All Control Sections

FO48%

AV35%

OT17%

Travis County

(OT = overturned, FO = fixed object, PD = pedalcyclist, PC = parked car, AV = another vehicle in transport)

Figure 5-7. First Harmful Event.

5-10

B69%

A25%

K6%


A30%

K9%

B61%

Texas Accidents

The primary accident type in Angelina County was recorded as colliding with another vehicle;more than half the crashes were of this type (see Figure 5-7). A third of the crashes were fixed-object crashes, and 11 percent involved overturned vehicles. Travis County crashes were dividedinto the same three primary categories seen for Angelina County, with fixed-object crashesaccounting for almost half of the total. All of the crashes in Martin County were collisionsbetween two vehicles. Almost half of the crashes in El Paso County involved fixed objects, withanother 38 percent involving collisions between two vehicles. The vast majority (96 percent) ofcrashes on the selected control sections are in three categories: another vehicle in transport, fixedobject, or overturned. Half of the crashes are collisions with another vehicle. The crashes for allTexas low-volume, rural two-lane roads are much more evenly distributed, although 93 percentof them are still in the same three categories as the study sections (see Figure 5-7). Based onthose observations it appears that western counties need to emphasize intersection treatments at asimilar level as roadway segment treatments, while eastern counties emphasize segmenttreatments over intersection treatments.

The level of injuries were similar between the selected roadways and the state. Between 60 and80 percent were non-incapacitating. Figure 5-8 illustrates the distribution of injuries. Crashes inthe eastern counties occurred on a variety of curves; however, well over half (65 percent) were onsections of roadway with no curve (see Figure 5-9). Western county crashes were predominantlyon straight sections of roadway; one crash was on a severe curve, and one was on a section ofunknown curvature. The distribution of crashes statewide is similar to that of the easterncounties, with slightly more than half occurring at locations with no curve. This observationindicates that eastern counties should continue their emphasis on addressing safety needs onhorizontal curves.

(K = Fatal, A = Incapacitating, B = Non-incapacitating)

Figure 5-8. Injury Severity Level.

5-11

Unknown4%

>6.013%

NC61%

0.1-1.95%

2.0-3.99%

4.0-5.98%

0.1-1.98%

2.0-3.98%

4.0-5.95%

NC65%

Unknown1%>6.0

13%

Eastern Counties

NC84%

>6.08%

Unknown8%

Western Counties

All Texas Counties

Figure 5-9. Crashes by Degree of Curvature (NC = no curve).

5-12

Dark - NSL35%

Dark - SL6%

Day59%


Dawn/Dusk2%

Dark - NSL37%

Dark - SL3%

Day58%

Texas Accidents

Over half of all control section crashes occurred in daylight hours, reflecting the trend in eachindividual county except for Martin County (see Table 5-3) which had all four of the crashesoccurring during the day. All Texas crashes exhibit a trend similar to that of the control sections,with a little more than half occurring during the day (see Figure 5-10).

(SL = street lights, NSL = no street lights)

Figure 5-10. Light Conditions.

Site Characteristics Data

The site characteristics data collected from field visits were entered into a spreadsheet for furtherexamination. The sections with more than two lanes were eliminated along with sections withADTs much greater than 2000. Those sections near 2000 ADT were retained. The sitecharacteristics data shown in Table 5-4 include the portions of the control section that have twolanes divided into the subsections that most closely approximated the traffic counting stationswith less than 2000 ADT.

Table 5-4 contains the site characteristics data for the control sections used as study sites,grouped by county. Total length is the combined length of all control sections from the countyincluded in the analysis. The average roadside environment score is based on a five-point scale,used for the area within 2 ft and within 10 ft of the paved surface. The scores were assignedbased on the most severe obstacle in the area, with values as follows:

1 = No fixed objects within 2 (or 10) ft of the edge of the paved surface2 = Yielding objects only (i.e., mailboxes, fence posts, delineators, etc.)3 = Combination of yielding and isolated rigid objects4 = Isolated rigid objects only (i.e., utility poles or trees more than 6 inches in diameter)5 = Many or continuous rigid objects (i.e., tree line, guardrail, stone fence, etc.)

5-13

Angelina Martin El Paso Travis25.4 33.8 22.6 21.71.2 0.4 0.1 2.79.0 11.5 10.0 11.6

10.7 12.1 12.8 13.2

Dir 1 < 2 ft 1.4 1.0 3.0 5.0Dir 1 < 10 ft 2.6 1.3 5.0 5.0Dir 2 < 2 ft 1.4 1.0 4.5 5.0Dir 2 < 10 ft 2.4 1.3 5.0 5.0

Dir 1 Trees Farmland Farmland TreesDir 2 Trees Farmland Farmland Trees

Dir 1 9.8 2.3 4.9 4.7Dir 2 8.7 2.8 3.3 8.2Total 18.5 5.1 8.2 12.9

Dir 1 1.4 0.8 1.3 0.6Dir 2 1.7 0.9 0.7 0.8Total 3.0 1.7 2.0 1.3

Dir 1 11.2 3.1 6.2 5.2Dir 2 10.4 3.7 4.0 9.0Total 21.6 6.9 10.2 14.2

Dir 1 0.0 5.5 4.5 1.0Dir 2 0.0 5.7 4.8 1.2

18.0 31.8 26.6 24.823.0 41.1 36.9 30.5

0 0 0 0

0 0 0.0 0.10.0 0 0 00.5 0.1 0.3 0.040 70 35 4555 70 55 5513 0 7 31

0.5 0 0.3 1.420 None 10 2050 None 35 50

Geometric Values

Traffic Control Values

Average Roadside Environment Score (scale of 1 to 5, with 1=no fixed objects, 5=many or continuous rigid objects)

Predominant Roadside Development

Driveway Density (driveways per mile)

Roadway Density (intersections per mile)

Access Density (driveway density + roadway density)

Average Shoulder Width (ft)

Number of Signals per mileNumber of RR Crossings per mileNumber of Stop Signs per mile

Advisory Speed Limits per mileMin Advisory Speed Limit (mph)Max Advisory Speed Limit (mph)

Min Posted Speed Limit (mph)Max Posted Speed Limit (mph)Number of Advisory Speed Limits

Min Pavement Width (ft)Max Pavement Width (ft)Median Width

Total Length (mi)Number of Vertical Curves per mileMin Lane Width (ft)Max Lane Width (ft)

Table 5-4. Site Characteristics for Control Sections Used as Study Sites, Grouped by County.

5-14

Predominant roadside development was determined by the technicians during their drive-throughs; categories included residential, commercial, farmland, trees, and park/school/campus. Lane and shoulder widths were measured in the field, from line to line for each lane and fromedgeline to edge of paved surface for each shoulder. The number of posted advisory speeds wasused as a surrogate for counting horizontal curves; the more advisory speeds and the lower theirvalues, the more winding the road was.

Table 5-4 shows the following patterns between the eastern counties and the western counties: • The number of vertical curves per mile are much higher in the eastern counties (1.2 to 2.7

vertical curves/mi) than the western counties (0.1 to 0.4 vertical curves/mi). • The average roadside environment score, particularly within 2 ft of the roadway has a similar

trend—the eastern counties (1.4 to 5.0) have a higher roadside environment score than thewestern counties (1.0 to 3.0). A roadside environment score of 1 is associated with no fixedobjects and a 5 represents many or continuous rigid objects.

• The observed roadside development is quite different between east and west, with farmlandbeing predominant in the west and trees in the east. (See Figure 5-2)

• Access density is also very different between east (14.2 to 21.6 access points/mi) and west (6.9to 10.2 access points/mi), especially when considering only driveway density (12.9 to 18.5driveways/mi in the east versus 5.1 to 8.2 driveways/mi in the west).

• Shoulders were much wider, on average, in western counties (4.5 to 5.7 ft) than easterncounties (0.0 to 1.2 ft) as were total pavement widths (26.6 to 41.1 ft in the west versus 18.0 to30.5 ft in the east).

• The number of advisory speeds posted on the study sites were much higher in the east (44)than in the west (7).

Relation of Crashes to Characteristics

Using the observed trends in the crash data and the characteristics data, in general, sites with ahigher crash rate have more vertical curves, more horizontal curves, more narrow lanes and/orshoulders, higher access density, a higher average roadside environment score, and a roadsidedevelopment that can more easily restrict sight distance and that may be more difficult to clearfrom the roadside.

As an example, Angelina County had the highest crash rate and the highest number ofintersection crashes of the four counties studied. Sections in Angelina County had the mostnarrow lane widths, no shoulders, and the highest access densities (driveway, roadway, andcombined). Conversely, Martin County had the lowest crash rate of the four counties; MartinCounty sections had the widest lanes and shoulders, the lowest access densities, the lowestnumber of vertical curves per mile, and no advisory speeds for horizontal curves.

SUMMARY OF FINDINGS

In the four counties studied for this task, two (Angelina and Travis) were counties with highKAB rates, and two (Martin and El Paso) were counties with lower KAB rates and that had asimilar number of miles of low-volume, rural two-lane highway. Angelina County had the

5-15

highest KAB crash rate as well as the highest number of crashes overall. Among the fourcounties, there is a distinct difference between eastern counties and western counties. Theeastern counties had the higher crash totals and rates, and they contained all of the fatal crashesconsidered in this study. In general, sites in the eastern counties had less driver-friendlycharacteristics, with more horizontal and vertical curves, narrower lanes and/or shoulders, lessforgiving roadside development, higher access density, and higher roadside development scores. Eastern counties also had more crashes at intersections than western counties.

6-1

CHAPTER 6

SUMMARY AND CONCLUSIONS

The state of Texas maintains nearly 80,000 centerline-miles of paved roadways serving about400 million vehicle miles per day. Over 62 percent of the centerline-miles are rural two-laneroads that, on average, have less than 2000 vehicles per day. These low-volume rural roadwayscarry less than 8 percent of the total vehicle miles on state-maintained (or on-system) highwaysand have approximately 11 percent of the total on-system vehicle crashes. These roadways alsohave relatively more severe injuries when vehicle crashes do occur. For example in 1999, about26 percent of the Texas on-system crashes are KAB crashes (i.e., fatal, incapacitating injury, andnon-incapacitating injury), while over 40 percent of the crashes on low-volume on-system roadsin 1999 were KAB crashes.

A project using the Texas DPS crash and TxDOT roadway inventory databases was conducted toidentify the common types of crashes on low-volume, rural two-lane highways. In 1999, therewere 45.7 KAB/100 MVMT on low-volume, rural two-lane highways. For all on-system roads,the rate was 31.5 KAB/100 MVMT. Therefore, a driver on a low-volume, rural two-lanehighway is more likely to be involved in a crash than the average for all Texas roads. The dataalso showed that a driver is more likely to be in a crash on a two-lane highway than on a multi-lane highway. In general, crashes on low-volume, rural two-lane highways occur betweenintersections, by a single vehicle running off the road. Crashes on curves and in dark, non-lightconditions are more common on low-volume, rural two-lane highways than on urban roads.

The project also demonstrated that more KAB crashes occurred in eastern counties than westerncounties. A sample of counties was selected to investigate which regional characteristics areassociated with high- and low-crash rates. The two counties within Texas with the highestaverage KAB rates for 1992 to 1998 were identified. Both of these counties were located in theeastern portion of the state. They were matched with counties in the western portion of the statethat had a similar number of low-volume, rural two-lane highway miles. The eastern countieshad the higher crash totals and rates, and they contained all of the fatal crashes considered in thisstudy. In general, sites in the eastern counties had less driver-friendly characteristics, with morehorizontal and vertical curves, narrower lanes and/or shoulders, less forgiving roadsidedevelopment, higher access density, and higher roadside development scores. Eastern countiesalso had more crashes at intersections than western counties.

Based upon the findings from the comparison of the crashes, the following are key directions adistrict may want to pursue when considering various types of low-cost improvements:

• treatments that either decrease the number of vehicles from leaving the roadway, especiallyon tight horizontal curves, or that better communicate the nature of the curve;

• improvements to reduce the number of nighttime crashes;

6-2

• treatments that reduce crashes at driveways; and• improvements to minimize severity of crashes if a vehicle leaves the road.

To obtain information about the types of treatments being used on these types of facilities, amailout survey was distributed to the TxDOT districts and to other states. Notable differencesbetween Texas responses and responses from other states include the following:

• For clear zone treatments, Texas had a much higher percentage for making culvertstraversable than other states.

• For wildlife control, other states use signs and reflectors more than Texas.• For additional lane improvements, other states use passing lanes and climbing lanes more

than Texas.• For pavement surface treatments, other states use centerline, edgeline, and intersection

rumble strips more than Texas.• For pavement markings, Texas uses thicker thermoplastic pavement markings and raised

pavement markings more than other states.• For sign improvements, other states use more advance intersection signs, diamond chevrons

signs, beacons on stop signs, and in-rail reflectors than Texas.• For other types of treatments, other states use more illumination than Texas.

The most frequently used treatments in Texas from all categories include:

• safety appurtenances,• mowing,• traversable culverts,• delineators,• advance curve signing,• advance stop signs,• raised pavement markers,• tree removal,• headwall removal,• advance intersection signage,• side slope flattening,• pavement edge maintenance,• pavement marking reapplication,• additional pavement markings, and• clear zone improvements.

The appendices of this report provide summaries of the effectiveness of various treatments asidentified from the literature.

R-1

REFERENCES

1. Fitzpatrick, K., D. W. Harwood, I. B. Anderson, and K. Balke. Accident MitigationGuide for Congested Rural Two-Lane Highways. NCHRP Report 440. TransportationResearch Board. (2000).

2. Krammes, Raymond A. “Interactive Highway Safety Design Model: A Brief Overview.”ITE National Conference. Kissimmee, FL (March 28-31, 1999).

3. Stamatiadis, N., S. Jones, and L. Aultman-Mall. “Causal Factors for Accidents onSoutheastern Low-Volume Rural Roads.” Transportation Research Record 1652. pp.111-117. (1999).

4. Fitzpatrick, K., D. W. Harwood, K. Bauer, and I. B. Anderson. Accidents on CongestedRural Two-Lane Highways. (Unpublished) Final Report NCHRP 15-17. TransportationResearch Board. (1999).

5. Roadway Safety Foundation. Roadway Safety Guide for Local Decision Makers andCommunity Leaders. Federal Highway Administration. (2001).

6. Harwood, D. W., and C. J. Hoban. Low-Cost Methods for Improving Traffic Operationson Two-Lane Roads: Informational Guide. Federal Highway Administration, FHWA-IP-87-2. (January 1987).

7. Deeter, D., and C. E. Bland. Technology in Rural Transportation. “Simple Solutions.”Federal Highway Administration, FHWA-RD-97-108. (October 1997).

8. Harwood, D. W., F. M. Council, E. Hauer, W. E. Hughes, and A. Vogt. Prediction of theExpected Safety Performance of Rural Two-Lane Highways. Federal HighwayAdministration, FHWA-RD-99-207. (2001).

9. Ivey, D. L., and L. I. Griffin. Safety Improvements for Low Volume Rural Roads, ReportNo. FHWA/TX-90/1130-2F. Federal Highway Administration, Washington, D.C. (1992).

10. Griffin, L. I., and D. L. Ivey. Safety Improvements for Low-Volume Rural Roads. Apaper submitted to the Transportation Research Board for the 71st Annual Meeting inJanuary 1992, Washington, D.C. (1991).

11. Maher, M. J., and I. A. Summersgill. Comprehensive Methodology for the Fitting ofPredictive Accident Models. Accident Analysis & Prevention, 28(3): 281-296 (1996).

12. Miaou, S.-P., Measuring the Goodness-of-Fit of Accident Prediction Models, FHWA-RD-96-040, Federal Highway Administration, U.S. Department of Transportation.(1996).

R-2

13. Zegeer, C.V., J. Hummer, D. Reinfurt, L. Herf, and W. Hunter. Safety Effects of Cross-Section Design for Two-Lane Roads – Volumes I and II. Report No. FHWA/RD-87/008. Federal Highway Administration, Washington, D.C. (1987).

APPENDIX A: LANE DEPARTURE

A-1

For low-volume, two-lane rural highways in Texas, about 67 percent of crashes occur at non-intersection locations. The top three categories used to indicate the first harmful event for a crashon these types of roadways were: fixed object (32 percent), other vehicle (32 percent), andoverturned (28 percent). The statistics for fixed object and overturned categories support theneed for keeping the vehicle on the roadway. A technique available for warning a driver that thevehicle is leaving the lane is rumble strips. A technique for helping to keep a vehicle on theroadway is to provide wider shoulders (see Appendix C). Rumble strips are raised or groovedpatterns placed on the pavement surface of a roadway or travel way. Their purpose is to providemotorists with an audible and vibrational warning that their vehicle has partially or completelyleft the lane.

RUMBLE STRIPS

As reported on one of their websites, a primary goal for the Federal Highway Administration is toreduce the number and severity of single-vehicle, run-off-road (ROR) crashes while preservingsafe use of the roadway by bicyclists and pedestrians. Roadway improvements that address therun-off-road issue include better geometric design, increased skid resistant roadway surfaces,more durable pavement markings, and more visible roadside signs. In recent years, several statetransportation agencies and toll road authorities have also installed and evaluated the effects ofshoulder rumble strips (SRS) on off-road crashes, particularly on rural freeways. The results ofthese evaluations have been overwhelmingly positive. FHWA has issued a Technical Advisory toprovide information on the state-of-the-practice for the design and installation of shoulder rumblestrips and to encourage their use on appropriate rural segments of the National Highway System(1). The Technical Advisory provides guidance on the design of the rumble strips andconsiderations for where to install them. Their recommendations on using rumble strips on roadsother than freeways in rural areas follow:

Because there are a significant number of ROR crashes on non-freeway facilities such asrural multilane and two-lane roadways, the FHWA recommends the use of shoulderrumble strips on those roadways for which an engineering study suggests that the numberof these crashes would likely be reduced by the installation of shoulder rumble strips. Insome cases, countermeasures, such as improved signing and markings, increasedpavement skid resistance, or other roadway improvements, may be more appropriate thanrumble strips or used in conjunction with them. When rumble strips are recommended,the following guidelines should be followed to the maximum extent practical:

• Standard milled rumble strips, installed as close to the edgeline as practical, should beused when an 8-ft (2.4 m) clear shoulder width remains available after installation of therumble strip.

• A modified design should be used when the remaining available clear shoulder width isless than 6 ft (1.8 m) wide and the road is used by cyclists. The most recent studiesindicate a milled depth of approximately 3/8 inch (10 mm) provides reasonable warningto most motorists while not being unduly dangerous to cross on a bicycle when necessary.Some states have also used narrower strips (i.e., less than 16 inches [400 mm]perpendicular to the direction of traffic) successfully. Others, as noted above, have


A-2

adopted a gap spacing to allow a cyclist to cross into the travel lane and back withouthaving to cross directly over the rumble strips.

• Rumble strips should not normally be used when their installation would leave a clearshoulder pathway less than 4 ft (1.2 m) wide for bicycle use.

The use of edgeline rumble strips has greatly increased in recent years. Several documentsdemonstrate their effectiveness or investigate alternative designs for bicyclists or other concerns. The use of centerline rumble strips on two-lane rural highways has also been implemented atcertain locations. Following is a summary of information available on edgeline and centerlinerumble strips.

EDGELINE RUMBLE STRIPS

Rumble strips warn motorists that they are leaving or about to leave the lane. Specific concernsthat affect the design of rumble strips and the locations where rumble strip installation isappropriate include: placement, weather, degradation of the pavement, type of pavement,pavement thickness, pavement overlay, noise, maintenance, motorist concerns, bicyclistconcerns, motorcyclist concerns, and potential for increase in head-on crashes caused by driversoverreacting to the edgeline rumble strip on a two-lane highway.

Shoulder rumble strips have proven to be an effective way of warning drivers that they areleaving, or about to leave, the roadway. Studies have shown that shoulder rumble strips areeffective against ROR/fixed object and ROR/rollover type crashes. Nationwide, ROR crashesaccount for approximately one-third of all traffic fatalities, with about two-thirds of these RORfatalities occurring in rural areas. It has been estimated that 40 to 60 percent of these crashes aredue to driver fatigue, drowsiness, or inattention (2).

Research has shown that shoulder rumble strips are an effective countermeasure to reduce ROR. Following is a summary of some of the findings:

• Rumble strips are estimated to reduce the rate of ROR crashes between 15 and 70 percent onthe FHWA website Rumble Strips (3).

• Data from the New Jersey Turnpike shows a 34 percent drop in ROR-type crashes afterinstalling shoulder rumble strips — at a time when overall crash rates increased by more than11 percent (3).

• The Pennsylvania Turnpike also saw a decrease in ROR crashes on their multilane facilitiesas reported in a 1997 publication. Reductions of about 100 crashes per year are attributed totheir rumble strips (4).

• Caltrans conducted an evaluation of the safety effects of continuous rumble strips on asphaltshoulders (5) for seven projects representing approximately 135 mi (217.4 km) of ruralfreeway in desert regions. The locations were described as having extremely monotonousdriving conditions. The ROR crash rate was reduced by 49 percent in the year followinginstallation.

• A 1985 FHWA study (6) included a detailed analysis of 10 sites. ROR crashes decreased by20 percent while rates on comparable control sites increased by 9 percent.


A-3

• A 1999 FHWA analysis (7) presented the results for two separate studies of continuousshoulder rumble strips (CSRS). The Illinois study involved comparisons of 55 treatment sitesto 55 control sites. This study showed an 18.3 percent reduction in single-vehicle, run-off-road crashes and a 13 percent reduction in single-vehicle, run-off-road injury crashes. Also,Illinois rural roadway data were collected and showed a reduction of 21.1 percent in single-vehicle, run-off-road crashes. Another study conducted in California posted results showinga 7.3 percent reduction in single-vehicle, run-off-road crashes, but it included a 13.4 percentstandard deviation implying insignificant results. This report also presented evidenceshowing a 23.6 percent increase in crashes among fatigued and drowsy drivers, but this tooincluded a high standard deviation (20.6 percent).

• A 1998 study (8) compared total run-off-road crashes before and after the installation ofCSRS and produced substantial results (see Table A-1).

• A 1999 study by Griffith (9) extracted data from California and Illinois and estimated thesafety effects of continuous rolled SRS on freeways. The results from the analysis estimatedthat continuous SRS reduced single-vehicle ROR crashes on average by 18.3 percent on allfreeways (with no regard to urban/rural classification) and 21.1 percent on rural freeways.

FHWA provided the data in Table A-2 as a summary of the associated crash reductions forshoulder rumble strips.

Table A-1. Before and After Data for Rumble Strips in New York State (8).

Year Total ROR Crashes Total Injuries Total Fatalities Million Vehicle Miles Traveled

Before and During Rumble Strip Installation

199119921993

557566588

358407328

17178

674476127792

After Rumble Strip Installation Completed

19961997

161 [74]*74 [88]

113 [72]54 [87]

4 [75]1 [95]

85128692

* Numbers in [ ] represent percent reduction in crashes from 1991

Table A-2. Shoulder Rumble Strips Studies and Associated Crash Reductions (2).

State (date) Roadway Type Percent Crash Reduction

Massachusetts (1997)New Jersey (1995)Washington (1991)Kansas (1991)FHWA (1985) – includes Arizona, California, Mississippi,Nevada, and North Carolina

Turnpike, RuralTurnpike, RuralSix LocationsTurnpike, RuralFive States, Rural

4234183420


A-4

Studies have also examined the effects on bicyclists for various shoulder rumble strips designs. FHWA (2) provides the following summaries for two recent studies:

• Moeur tested 28 bicyclists (5 basic, 17 skilled, and 6 experienced) in a 2000 Arizona fieldstudy by having them ride over various skipped SRS sections to determine acceptable skippatterns. It was determined that 12-ft (3.7 m) skips in ground-in SRS pattern wouldacceptably permit bicyclists to cross at high speeds (speeds were assumed to be between 23 to28 mph or 37 and 45 km/h). Either 40- or 60-ft (12.2 or 18.3 m) cycles for the skip patternwere determined acceptable.

• The objective of a Pennsylvania project performed in 2000 was to develop new SRSconfigurations that decrease the level of vibration experienced by bicyclists while providingan adequate amount of stimulus to alert inattentive or drowsy drivers. Six configurationswere tested by 25 intermediate and advanced bicyclists. The researchers recommended theadoption of two new bicycle-tolerable rumble patterns, one for non-freeway facilitiesoperating near 55 mph (88 km/h) and the other for those operating at 45 mph (72 km/h).

• In 2001, the California Department of Transportation performed a study of various SRSdesigns. Six test vehicles, ranging from a compact automobile to large commercial vehicles,were used to collect auditory and vibrational data while traversing the SRS. Fifty-fivebicyclists of various skill levels and ages evaluated the SRS designs. The recommendation ofthe study was to replace the existing rolled SRS design with a milled SRS design that is 1 ft(300 mm) in transverse width and 5/16 ± 1/16 inches (8 ± 1.5 mm) in depth on shoulders thatare at least 5 ft (1.5 m) wide. For shoulders less than this width, the installation ofraised/inverted profile thermoplastic was recommended.

• A 2001 study in Colorado developed recommendations based on the input of 29 bicyclists aswell as vibrational and auditory data collected in four different types of vehicles. Of the 10styles tested, those that provided the most noticeable vibrational and auditory stimuli to thevehicle were rated worst by bicyclists.

The benefit-cost ratio for rumble strips has been estimated in two recent studies. A 1999 report(9) offers benefit-cost assumptions based on the reduction of crashes and the total cost of a singlerun-off-road crash. In this comparison, it is estimated that an average cost of $62,200 isprevented every three years based on an investment of $217. A 1998 analysis (8) compared theestimated cost of rumble strips including installation, maintenance, and protection of traffic andthe cost of fatal, non-incapacitating, and property damage crashes on 486 mi (783 km) of NewYork Thruway. The results, based on reduction of crashes and a six year estimated maintenance-free life, showed a benefit/cost ratio of 182 and a yearly savings of $58,893,500.

CENTERLINE AND EDGELINE RUMBLE STRIPS

In addition to using edgeline rumble strips on two-lane rural highways, some states are alsoinstalling centerline rumble strips as a treatment to reduce the number of head-on collisions. AFHWA Technical Advisory (1) stated that “some states have installed milled centerline rumblestrips on two-lane roads having a history of head-on and opposite-direction sideswipe crashes.Most of these installations have consisted of transverse grooves extending across the doubleyellow centerline and the space between them. Initial evaluation efforts have shown reductions in


A-5

Figure A-1. Example of Shoulder and CenterlineRumble Strips.

the types of crashes that centerline rumble strips address.” FHWA is seeking candidate sites for a nationwide study on the treatment. NCHRP Report 440 (10) reported on three sites where bothedgeline and centerline rumble strips are being used. A summary of those three sites follows.

Centerline Rumble Strips and Inverted Profile Thermoplastic Edgelines for a CaliforniaState Route

An increase in the number of fatal crashes on a state route in California in 1995 generatedconcerns from the local community and elected officials. In the previous nine years, the averagenumber of fatal crashes was 2.7 per year. In 1995, the number of fatal crashes was six with atotal of 14 people killed. Financial constraints limited the options for improvements such asexpanding the highway to a four-lane divided roadway or widening the shoulders. Caltranslooked for improvements that could correct driver behavior in a manner that would reduce thefatal head-on crashes. The following were installed as part of the project:

• Double yellow stripes were replaced with a rumble strip (to provide an audible and vibratorywarning) and raised profile thermoplastic traffic striping (to enhance nighttime visibility andprovide both audible and vibratory warning for straying motorists). In addition, yellow retro-reflective pavement markers were also installed between the rumble strip and raised profilethermoplastic. The spacing between the double yellow stripes was increased to 28 inches(71.1 cm) to accommodate the 16 inches (40.6 cm) used for the ground-in rumble strip, the 4inches (10.1 cm) for each thermoplastic stripe, and the yellow retro-reflective markers. Figure A-1 shows a sample of the treatment.

• Solid yellow centerline stripes in one direction (no passing sections) were replaced withraised profile thermoplastic striping.

• Markers were added to the center of the roadway in passing sections to provide audiblewarning to motorists crossing the center of the road.

• Inverted profile thermoplastic striping with rumble strips was placed on shoulders with aminimum width of 6 ft (1.8 m). Shoulder widths less than 6 ft (1.8 m) received invertedprofile thermoplastic striping only. Aminimum of 4 ft (1.2 m) ofunobstructed shoulder was preserved toaccommodate bicycle traffic.

• Black non-reflective pavement markerswere placed in the intervals betweenstripes in the two-lane passing zones toincrease the rumble effect.

• The concrete bridges on the projectreceived raised profile and invertedprofile thermoplastic striping on thecenterline where double yellow stripingexisted. Inverted profile thermoplasticstriping was placed at the edge of thetravel way. No ground-in rumble stripswere constructed on the bridges.


A-6

The installation of the pavement treatments was completed in November 1996. The estimatedcost of the work in January 1996 was $789,000. Crash data for the roadway segment wasavailable for 25 months after the installation of the treatment. In addition, 34 months of beforedata were obtained. The 23.5-mi (37.8 km) segment experienced fewer crashes in the afterperiod than the before period. An average of 4.5 crashes occurred per month in the before periodand 1.9 crashes per month occurred in the after period. In the before period, 10 crashes resultedin fatalities while the after period only included one fatal crash. The distribution of crash types(e.g., head-on, rear-end, etc.) and primary cause (e.g., improper turn, speeding, etc.) remainedrelatively constant between the two periods. This limited crash review indicates that rumblestrips could be an effective treatment in reducing crashes.

Rumble Strips along with Other Treatments

A principal highway that connects an interstate to small towns has become a commuter andrecreational route. Recreational travel is expected to increase and is particularly pronounced onsummer, fall, and holiday weekends. Commuting is a relatively new use of the highway. Theadvent of the interstate and the growth of commercial development in the northern suburbs of amajor city have made the small towns a reasonable commute. As these towns develop asresidential communities, the peak-period commuter traffic demands placed on the road isexpected to increase. The large and increasing recreational and commuting traffic have focusedattention on the capacity and safety of the highway. Both the crash and severity rates for the highway are generally below average when compared tothe sections of roadway within the metropolitan area with similar design characteristics. However, when compared to statewide averages, the rates on the highway are slightly higher thanthe average. Table A-3 summarizes the three-year crash totals for 1991-1993.

Table A-3. Crash Totals for 1991-1993.

Crash Type Number% of TotalReported

Rear-endSideswipe – same directionApproach turnRight angleRan off road – left sideSideswipe – opposite directionHead-onRan off road – right sideType not reportedTOTAL

1162318491681130153424

43961863411--

100

The high percentage of rear-end crashes suggests that a disproportionately high level of directaccess is being provided by the highway, which is in conflict with the high overall travel speedson the roadway. The high level of direct access between adjacent land uses and the highway is aproduct of both the historic growth pattern of the corridor and the lack of parallel routes throughlarge segments of the corridor.


A-7

A number of short-term improvements were made to the highway to address safety concerns. These include the installation of traffic signals and channelization at two intersections, right-turnand right-turn/bypass lanes, one left-turn lane, raised pavement markers on horizontal curves, andcenterline and edgeline rumble strips.

The rumble strip work was completed in the fall of 1995 at a cost of $54,000. Approximately 15mi (24.2 km) were treated. The DOT completed a two-year before-and-after study of the crasheswithin the rumble strip areas. Table A-4 presents a summary of the before and after crasheslisted by crash type at two sections along the highway where rumble strips were installed. Whilesome crash types decreased, there does not appear to be a significant crash reduction attributed tothe installation of the rumble strips (e.g., most of the reduction in crashes occurred in the “other”category rather than in the “off road - right” category).

Table A-4. Crash Summary (Rumble Strip Installations).

Type of Crash Number of Crashes

Section 1 Section 6

Beforea Afterb Beforea Afterb

Rear-endSideswipe - passingSideswipe - opposingLeft turnOff road - leftOff road - rightRight angleHead-onOthers

13623789230

1661245329

7311132029

211102628

TOTAL 80 48 47 23

a 11/01/93 to 10/31/95 (24 months)b 11/01/95 to 10/31/97 (24 months)

Rumble Strips, Lane Striping, and Guardrail Installations

A highway that connects a major northwestern city and nearby suburbs and small cities serves asa high-volume commuter route. The roadway generally has 12-ft (3.7 m) lanes and 8-ft (2.4 m)(or greater) shoulders with signals and left-turn bays at selected intersections. Previousinvestigation into the crash characteristics of the roadway showed that several of the crashes wereopposite-direction crashes, which suggests that they were due to passing maneuvers. The DOTselected centerline rumble strips along with lane striping and guardrail installation as thecountermeasures due to the high number of opposite direction crashes. These treatments were


A-8

viewed as an interim measure until sufficient fundingwas available for widening the highway. The work wasperformed by the maintenance personnel for anapproximate 10-mi (16.1 km) section. Figure A-2shows a sample of the rumble strips.

The treatments were constructed between July 28, 1995,and September 29, 1995. The DOT provided crash datafor one year before and one year after the installation ofthe treatments. Table A-5 presents a summary of thebefore and after crashes listed by crash severity and typefor the section. A crash reduction of 23 percent wasexperienced between the two years of data with most ofthe crash reduction being from a decrease in rear-endcrashes.

Figure A-2. Example of Installed Rumble Strips.

Table A-5. Crash Summary (Rumble Strip Installation).

Type of Crash Number of Crashes

Beforea Afterb

Property Damage InjuryFatal

53462

45312

TOTAL 101 78

Alcohol RelatedFixed ObjectRear-EndOpposite DirectionEntering at AngleOverturnHazardous MaterialPedestrianOther

1125361645004

920251235013

TOTAL 101 78

a7/28/94 to 7/27/95 (12 months)b9/29/95 to 9/28/96 (12 months)


A-9

REFERENCES

1. Shoulder Rumble Strip Technical Advisory Online.http://safety.fhwa.dot.gov/fourthlevel/rumstrp_ta.htm

2. Shoulder Rumble Strip Synthesis Online.http://safety.fhwa.dot.gov/fourthlevel/exec_summary.htm

3. Rumble Strips Online.http://safety.fhwa.dot.gov/fourthlevel/pro_res_rumble.library.htm#Papers

4. Rumble Strips: A Sound Investment. Federal Highway Administration, Publication NumberFHWA-SA-99-017. (1999).

5. Hickey, J. J. “Shoulder Rumble Strip Effectiveness,” Transportation Research Record 1573,Transportation Research Board, Washington, D.C. (1997).

6. Chaudoin, J. H., and G. Nelson. Interstate Routes 15 and 40 Shoulder Rumble Strips. ReportNo. Caltrans-08-85-1, Traffic Operations Branch—District 8, California DOT. (August1985).

7. Griffith, M. Safety Evaluation of Continuous Shoulder Rumble Strips Installed on Freeways.HSIS Summary Report. (December 1999).

8. Perrillo, K. The Effectiveness and Use of Continuous Shoulder Rumble Strips. FederalHighway Administration. (August 1998).

9. Griffith, M. Safety Evaluation of Continuous Shoulder Rumble Strips Installed on Freeways.Transportation Research Record 1665. (1999) pp. 28-34.

10. Fitzpatrick, K., D. W. Harwood, I. B. Anderson, and K. Balke. Accident Mitigation Guide forCongested Rural Two-Lane Highways. NCHRP Report 440. Transportation Research Board,Washington, D.C. (2000).

http://safety.fhwa.dot.gov/fourthlevel/rumstrp_ta.htm

http://safety.fhwa.dot.gov/fourthlevel/exec_summary.htm

http://safety.fhwa.dot.gov/fourthlevel/pro_res_rumble.library.htm#Papers

APPENDIX B: ROADWAY DEPARTURE

B-1

Appendices A and C discuss treatments that can be used to minimize run-off-road crashes. Appendix A provides information on rumble strips, and Appendix C discusses the benefits ofwider shoulders. Once a vehicle has left the roadway, any obstacle located on the roadside hasthe potential for being hazardous to an errant vehicle; therefore, efforts should be made toremove, protect, or make forgiving an obstacle or object that has to be located in the right-of-way. In addition to crash frequency, the severity of crashes involving specific roadside obstaclesis also important.

A 1978 FHWA study by Perchonok et al. analyzed crash characteristics of single-vehicle crashes,including crash severity related to types of objects struck (1). For non-rollover fixed-objectcrashes, the obstacles associated with the highest percent of injury occurrences are, in order:bridge or overpass entrances, trees, field approaches (i.e., ditches created by driveway), culverts,embankments, and wooden utility poles. Actual percent injuries and fatalities of these crashesare shown in Table B-1. Obstacle types with the lowest crash severity include small sign posts,fences, and guardrails (1).

Table B-1. Severest Injury by Object Struck in Non-Rollover Crashes (1).Object Crash Sample

SizePercentInjured

PercentKilled

Bridge/Overpass EntranceTree

Field ApproachCulvert

EmbankmentWood Utility Pole

B/O SiderailRock(s)Ditch

GroundTrees/Brush

GuardrailFence

Small Sign Post

8866775231406598 82 73368153255284325 76

75.067.966.762.357.651.251.249.348.948.438.431.724.322.4

15.97.21.36.14.42.32.41.41.13.32.01.80.31.3

Total 3681 50.8 3.6

A separate analysis was also conducted for severity of crashes involving ditches. The authorsfound that ditches which were 3 ft (0.9 m) or deeper were associated with a higher percent ofinjury crashes (61 percent) when compared to crashes involving ditches 1 to 2 ft (0.3 to 0.6 m)deep (54 percent injury). Percent fatal crashes were about the same for each depth category (i.e.,about 5 percent for both the 1- and 2-ft [0.3 and 0.6 m] group and the 3-ft-plus [0.9 m plus]group).


B-2

RECOVERY DISTANCE

The concept of a forgiving roadside recognizes that motorists do run off the roadway and that atraversable recovery area could lessen serious crashes and injuries. Ideally, this recovery area or“clear zone” should be free of obstacles such as unyielding sign and luminary supports, non-traversable drainage structures, utility poles, and steep slopes. Design options for the treatmentof these features have been generally considered in the following order:

� Remove the obstacle or redesign it so it can be traversed safely.• Relocate the obstacle to a point where it is less likely to be struck.• Reduce the impact severity by using an appropriate breakaway device.• Redirect a vehicle by shielding the obstacle with a longitudinal traffic barrier and/or crash

cushion if it cannot be eliminated, relocated, or redesigned.• Delineate the obstacle if the above alternatives are not appropriate.

The roadside recovery distance is a relatively flat, unobstructed area adjacent to the travel lane(i.e., edgeline) where there is a reasonable chance for an off-road vehicle to safely recover (2). Therefore, it is the distance from the outside edge of the travel lane to the nearest rigid obstacle(e.g., bridge rail, tree, culvert, utility pole), steep slope, non-traversable ditch, or other threat(e.g., cliff, lake) to errant motor vehicles.

Maintaining an adequate recovery area, free of obstacles and obstructions, is one way of reducingthe crash exposure on two-lane congested roadways. Recommended roadside recovery distances(or clear zones) can be obtained from the Roadside Design Guide (3). The data were based onlimited empirical data that were then extrapolated to provide data for a wide range of conditions;therefore, the numbers obtained represent a “reasonable measure” of the degree of safetysuggested for a particular roadway.

Along a roadway section, the roadside recovery distance may vary considerably. The recoverydistance for a roadway section can be determined by taking an average of measurements (e.g., 3to 5 measurements per mi [2 to 3 per km] on each side of the road). Roadside recovery distancesof 0 to 30 ft (0 to 9.2 m) are generally recorded.

Examples of roadside improvements that can increase the recovery distance include cutting treesnear the roadway, relocating utility poles further from the road, and use of side slopes of about1:4 or flatter.

For roadways with limited recovery distances (particularly less than 10 or 15 ft [3.1 or 4.6 m]from the roadway edgeline) where roadside improvements are proposed, crash reduction factorsmay be found in Table B-2. For example, increasing the roadside recovery distance by 12 ft (3.7m) (e.g., from 4 to 16 ft [1.2 to 4.9 m]) will reduce “related” crashes (defined as including run-off-road, head-on, and sideswipe crashes) by an estimated 29 percent.


B-3

Table B-2. Crash Reduction Factors Due to Increasing Roadside Clear Recovery Distance (2).

Amount of Increased RoadsideRecovery Distance, m (ft)

Percent Reduction in RelatedCrash Types* (%)

1.5 (5)2.4 (8)

3.1 (10)3.7 (12)4.6 (15)6.2 (20)

132125293544

*Related crash types = run-off-road, head-on, and sideswipe

SIDE SLOPES

The steepness of the roadside slopes, or side slopes, is a cross-sectional feature that affects thelikelihood of an off-road vehicle rolling over or recovering back into the travel lane. Existingguidelines for acceptable side slopes have historically been based on computer simulations andobservations of controlled vehicle test runs on various slopes as well as on “informed”judgments.

Figure B-1 shows a relationship between single-vehicle crashes and field-measured side slopes. As shown in Figure B-1, single-vehicle crashes (as a ratio of crashes on a 1:7 slope) are highestfor slopes of 1:2 or steeper and drop only slightly for 1:3 slopes. Single-vehicle crashes thendrop linearly (and significantly) for flatter slopes. This plot represents the effect of side slopeafter controlling for ADT and roadway features (2).

0.9

1

1.1

1.2

1.3

1.4

1:2 orsteeper

1:3 1:4 1:5 1:6 1:7 orflatter

Side Slope Ratio

Rel

ativ

e S

V A

ccid

ent

Rat

e

.

Figure B-1. Plot of Single-Vehicle (SV) Crash Rate for a Given Side Slope to Single-Vehicle Crash Rate for a Side Slope of 1:7 or Flatter (2).


B-4

In fill sections, side slopes that are 1:4 or flatter are generally desirable. When side slopes are 1:4or flatter, motorists encroaching on the side slope can generally stop their vehicles or slow themenough to recover safely (as long as the appropriate recovery distance, free of obstacles, has beenprovided). Side slopes between 1:3 and 1:4 are traversable, but most motorists will be unable tostop or return to the roadway easily. In these sections, a runout area (see Figure B-2) may berequired at the toe of the non-recoverable slope of the recovery area. Side slopes that are greaterthan 1:3 are considered to be critical because a vehicle that leaves the roadway is likely tooverturn on such side slopes. If a side slope steeper than 1:3 begins closer to the traveled waythan the suggested clear zone distance, a barrier might be warranted if the slope cannot beflattened easily. See AASHTO’s Roadside Design Guide (3) for information on warrantingbarriers for side slopes.

Figure B-2. Example of a Side Slope Design in a Fill Section (3).

When a highway is in a cut section, the back slope may be traversable depending upon its relativesmoothness and the presence of fixed obstacles. If the slope between the roadway and the baseof the back slope is traversable (1:3 or flatter) and the back slope is obstacle-free, it may not be asignificant hazard, regardless of its distance from the roadway. On the other hand, a steep,rough-sided rock cut should normally begin outside the clear zone or be shielded. A rock cut isnormally considered to be rough-sided when the face will cause excessive vehicle snaggingrather than provide relatively smooth redirection. Warrants for the use of a roadside barrier inconjunction with cut slopes can be found in the AASHTO Roadside Design Guide.

Ditches represent a unique roadside hazard in many areas. Designed primarily to collect andconvey storm water runoff, their design should also consider what would happen if a vehiclewere to leave the roadway. The AASHTO Roadside Design Guide gives preferred fore slopesand back slopes for basic ditch configurations. Cross sections that fall in the shaded region ofeach of the figures are considered to be traversable by errant vehicles. Ditch sections that falloutside the shaded region are considered less desirable and their use should be limited wherehigh-angle encroachment can be expected, such as on the outside of a relatively sharp horizontalcurve. Types of improvements to ditches with cross sections needing improvements and locatedin vulnerable areas include the following: reshape the ditch to conform to a more “forgiving”


B-5

design, convert ditch to a closed drainage system using culverts and pipes, and shield ditch totraffic using a traffic barrier.

Figure B-1 shows the relationship used to develop crash reductions matching various side slopeflattening projects. The percent reductions are presented in Table B-3 for single-vehicle and totalcrashes. For example, flattening an existing 1:2 side slope to 1:6 should result in a reduction ofapproximately 21 percent and 12 percent of single-vehicle and total crashes, respectively (2). These reductions assume that the roadside slope to be flattened is relatively clear of rigidobstacles.

The use of flatter slopes not only reduces the crash rate, but it may also reduce rollover crashes,which are typically quite severe. In fact, injury data from three states reveal that 55 percent ofrun-off-road rollover crashes result in occupant injury, and 1 to 3 percent end in death. Of allother crash types, only pedestrian crashes and head-on crashes result in higher injury percentages(2). The FHWA study found that side slopes of 1:5 or flatter are needed to significantly reducethe incidence of rollover crashes (i.e., not 1:4, as is often assumed) (2).

Table B-3. Effect (%) of Side Slope Flattening on Single-Vehicle and Total Crashes (2).

Side SlopeBefore

Condition

Side Slope in After Condition

1:4 1:5 1:6 1:7 or Flatter

SingleVehicle Total

SingleVehicle Total

SingleVehicle Total

SingleVehicle Total

1:21:31:41:51:6

10%80--

65---

151460-

983--

21191260

121173-

272619148

15151185

Note: These values are only for two-lane rural roads.

ROADSIDE OBSTACLES

Trees

Trees become potential obstructions by virtue of their size and location in relation to vehiculartraffic. Generally, a single tree with a trunk diameter greater than 5.9 in (150 mm) is considereda fixed object. When trees or shrubs with multiple trunks or groups of small trees are together,they may be considered as having the effect of a single tree with their combined cross-sectionalarea. Tree removal should be considered when those trees are determined both to be obstructionsand to be in a location where they are likely to be hit. If tree removal is impractical or infeasible,then shielding the trees with some type of roadside barrier may be justified. The reader isreferred to AASHTO’s Roadside Design Guide (3) for more information about the warrantingand design of roadside barriers for protecting trees.


B-6

Tree crashes can be reduced based on crash reductions shown in Table B-4. For example,clearing trees by 10 ft (3.1 m) (e.g., from 8 to 18 ft [2.4 to 5.5 m]) will reduce tree crashes by anexpected 57 percent. These values assume that by clearing trees back from the roadway, run-off-road vehicles would have an additional roadside area to recover provided the trees were not on asteep side slope. Since trees are the fixed object most often struck on many rural roads, clearingtrees back from the road (particularly on roads with severe alignment) can be an effectiveroadside safety treatment (4).

Table B-4. Percent Reductions in Specific Types of Obstacle Crashes Due toClearing/Relocating Obstacles Further from the Roadway (4).

Increase in ObstacleDistance (IOD)* m

(ft)Trees(%)

Mailboxes,Culverts,

& Signs (%)Guardrails

(%)Fences/Gates

(%)0.9 (3)1.5 (5)2.4 (8)3.1 (10)4.0 (13)4.6 (15)

223449576671

14233440

N.F.*N.F.

36537078

N.F.N.F.

20304452

N.F.N.F.

*Notes:N.F. = generally not feasible to relocate obstacles to specified distances.IOD = amount of increase in obstacle distance from roadway.This table is appropriate only for obstacle distances of 30 ft (9.1 m) or less and only on two-lane rural roadways.

Culvert Headwalls

Drainage features should be designed and built with both hydraulic efficiency and roadside safetyin mind. Common drainage structures that might represent a hazard to motorists whose vehiclesleave the roadway include the following: curbs, parallel and transverse pipes and culverts, anddrop inlets.

The following list shows several options (in order of preference) to modifying drainagestructures:

• Eliminate non-essential drainage structures.• Design or modify drainage structures so they are traversable or present a minimal hazard to

an errant vehicle.• If a major drainage feature cannot effectively be redesigned or relocated, it should be shielded

by a suitable traffic barrier if it is in a vulnerable location.

AASHTO’s Roadside Design Guide (3) should be consulted for details on the design of thesestructures.


B-7

Culvert headwalls can result in serious injury or death when struck at moderate or high speeds onrural roadways. While relocating such culverts further from the roadway may be feasible undercertain conditions, the ideal solution would be to reconstruct the drainage facilities so that theyare flush with the roadside terrain and present no obstacle to motor vehicles. Such designs wouldessentially eliminate culvert crashes although run-off-road vehicles could still strike otherobstacles (e.g., trees) beyond the culverts or roll over on a steep side slope (see discussion of sideslope in an earlier section). Crash reductions which correspond to placement of culvertheadwalls further from the roadway are shown in Table B-4. For example, a 40 percentreduction in culvert hits is expected for culverts located 15 ft (4.6 m) from the road compared to5 ft (1.5 m) (i.e., a 10-ft (3.1 m) difference in distance) (4). Other useful information on drainagestructures is contained in the Roadside Design Guide (3).

Ross et al. (5) developed preliminary guidelines for minimum spacing of driveways on highspeed roadways (see Table B-5). The guidelines address safety concerns related to run-off-roadcrashes. The purpose of the guidelines is to minimize the risk to an errant motorist who leavesthe road, crosses a driveway/sloped-end culvert, and then becomes airborne. It is desirable tohave a safe recovery area downstream from the driveway — one that is free of hazardousfeatures, including another driveway.

Table B-5. Tentative Spacing Guidelines for Multiple Driveways (5).

Driveway Slope Speed, mph (km/h) Minimum Spacing Indicatedft (m)

1:6 45 (72.5)50 (80.5)55 (88.6)60 (96.6)

50.2 (15.3)75.1 (22.9)100.0 (30.5)100.0 (30.5)

1:8 45 (72.5)50 (80.5)55 (88.6)60 (96.6)

24.9 (7.6)24.9 (7.6)50.1 (15.3)75.1 (22.9)

1:10 45 (72.5)50 (80.5)55 (88.6)60 (96.6)

00

24.9 (7.6)24.9 (7.6)

Sign Support and Placement

Roadside signs can be divided into three main categories: overhead signs, large roadside signs,and small roadside signs. Each sign type requires a different hardware and safety treatment.Because overhead signs generally require massive support systems that cannot be madebreakaway, they should be installed on or relocated to nearby overpasses or other structures,where possible. If it is not possible to locate the supports outside the clear zone, overheadsupports are to be shielded with an appropriate barrier. Large roadside signs may be defined as


B-8

those greater than 53.8 ft2 (5 m2) in area. They typically have two or more support poles that canbe made breakaway. The basic concept of the breakaway support is to provide a structure thatwill resist wind and ice loads, yet fail in a safe and predictable manner if struck by a vehicle. Small roadside signs may be defined as those supported on one or more posts and having a signpanel area of less than 53.8 ft2 (5 m2). Although not perceived as a significant obstruction, smallsigns can cause significant damage to impacting automobiles. The most common methods formaking base supports for small roadside signs breakaway include the following: base bending oryielding sign supports, fracturing sign supports (e.g., wood, steel posts, or steel pipes connectedat ground level to a separate anchor), and slip base designs. AASHTO’s Roadside Design Guide(3) should be consulted on how to design breakaway supports for roadside signs.

Sign placement is largely a function of readability to drivers, so in some respects signs should notbe placed too far from the road. Even though sign posts represent a roadside obstacle, signplacement must be within the driver’s cone of vision to be useful. Where practical, the use ofbreakaway sign posts is highly desirable to minimize the severity of impacts between motorvehicles and the posts. Where not practical, the sign should be relocated further from thepavement edge. The percent reductions in sign crashes are given in Table B-4 for variousdistances of the signs from the roadway.

Mailboxes

AASHTO’s A Guide for Erecting Mailboxes on Highways (generally called the Mailbox Guide)(6) contains information on mailbox supports and their location on the roadside. The followingguidelines should be used for installing mailbox supports:

• Mailbox supports, which should be considered as nominal, are 3.9 inches by 3.9 inches (100mm by 100 mm) or 3.9 inches (100 mm) diameter wood posts, or a metal post with a strengthno greater than a 2.0 inches (50 mm) diameter standard strength steel pipe, embedded nomore than 23.6 inches (600 mm) into the ground. For example, a single 0.4-lb/ft (3.0 kg/m) U-channel support would be acceptable under this structural limitation. Mailbox supportsshould not be set in concrete unless the support design has been shown to be safe by crashtests.

• Mailbox-to-post attachments should ideally prevent mailboxes from separating from theirsupports under vehicle impacts. The Mailbox Guide contains information on attachmentsthat prevent their separation (6).

• Multiple mailbox installations should meet the same criteria as single mailbox installations. Multiple support installations should have their supports separated a minimum distance equalto three-fourths of their heights above ground. This will reduce interaction between adjacentmailboxes and supports.

While relocating mailboxes further from the road would be expected to reduce the frequency ofmailbox crashes, such relocation is not practical in many situations. A more promisingalternative, which would affect crash severity but not crash occurrence, would be to make use ofmailboxes


B-9

with less rigid posts or breakaway design in place of the heavy steel, wooden posts, or multipleposts (4). Recent research has documented the injury reduction from breakaway mailbox posts(7).

Guardrail

Guardrail is installed along roadways to shield a vehicle from striking a more rigid obstacle orfrom rolling down a steep embankment. When installed, guardrail is generally positioned at thegreatest practical distance from the roadway to reduce the incidence of guardrail impacts. Thus,it is not often feasible to relocate guardrail further from the roadway along a section unless someflattening of the roadside occurs. However, when it is feasible to flatten roadsides to a relativelymild slope (e.g., 1:5 or flatter) with appropriate removal of obstacles, then guardrail should beremoved since the guardrail itself presents an obstacle which vehicles can strike. The crashreductions in Table B-4 for guardrail placement illustrate the crash benefits from relocatingguardrail (4).

Russell et al. (8) developed guidelines for the use of guardrail on low-volume roads in Kansas. Low-volume roads were defined as roads with �400 ADT. The Kansas Department ofTransportation (KDOT) wanted guidelines for using guardrail on low-volume roads in Kansasbased on a cost-effectiveness analysis from adapting the microcomputer program ROADSIDE. KDOT was only interested in guidelines for three types of roadside obstacle: 1) culvert-straightwings; 2) culvert-flared wings; and 3) culvert-pipe/headwall. Conditions considered were offsetdistance, ADT, speed, and ditch depth.

Fences and Gates

Fences and gates are sometimes placed by private property owners just beyond the highway right-of-way and can present a hazard to run-off-road vehicles. As shown in Table B-4, the effect ofrelocating fences is a 20 percent crash reduction for 3 ft (0.9 m) of relocation, 44 percent for 8 ft(2.4 m) of relocation, and 52 percent for 10 ft (3.1 m) of relocation. Unfortunately, having fencesrelocated further from the roadway could require that an agency purchase more right-of-wayalong a route, which could be quite expensive (4).

Utility Poles

Utility poles represent a significant hazard on two-lane rural highways. Significant reduction incrashes on two-lane rural roadways can be achieved by reducing, eliminating, or protecting utilitypoles. Motor vehicle collisions with utility poles result in approximately 10 percent of all fixed-object fatal crashes annually. The frequency and severity of crashes with utility poles is affectedby three factors:

• the number of utility poles per mile (pole density),• the proximity of the poles to the edge of the travel way (pole offset), and• the nature of their design (i.e., whether they are breakaway or unyielding).


B-10

Because most utility poles are generally privately owned and only installed on publicly ownedright-of-way, they are often not under the direct control of the public agency, complicating the implementation of effective countermeasures.

Several options exist for reducing crash frequency and severity with existing utility poles,including the following:

� placing utility lines underground,� increasing the lateral pole offset,� increasing pole spacing,� installing breakaway utility poles, and� placing barriers around utility poles. A site-specific benefit/cost analysis can be performed to decide which corrective measure is mostcost-effective at a specific high crash-frequency location.

Placing Utility Lines Underground

This countermeasure involves removing the utility poles and burying the utility linesunderground. Theoretically, placing the utility lines underground should eliminate 100 percentof the run-off-road/fixed-object crashes where the first harmful event is striking a utility pole. However, the true level of effectiveness of placing the utilities underground will depend on thenumber and proximity of other fixed objects in the roadway clear zone. Overall crash frequencyis likely to be unaffected by placing utilities underground because vehicles are likely to bestriking other fixed objects in the right-of-way. In one study, fatal crashes were reduced 38percent as a result of burying the utilities underground; however, injury crashes increased 1.5percent (6).

Increasing Lateral Pole Offset

As with other fixed objects in the right-of-way, the most desirable solution for correcting crashesis to locate the poles where they are least likely to be struck. This countermeasure is aimed atreducing utility pole crashes by increasing the distance a vehicle has to travel before striking autility pole. Relocating poles farther from the roadway will generally reduce the frequency ofutility pole crashes. Table B-6 (9) shows the reduction in utility pole crash frequency as a resultof increasing the offset of utility poles from the pavement edge. There is no conclusive evidenceto support that pole relocation will have a significant effect on the severity of utility pole crashes.

Reducing the Number of Poles

Utility pole spacing varies widely based on the type of lines. For telephone and small electriclines, pole spacing generally ranges about 100 ft (31 m). For larger voltage power lines (morethan 69 KV), spacings are commonly 500 ft (152 m) apart or more. One way to reduce crashfrequency is to reduce the number of poles (or pole density) in a given section of roadway.


B-11

There are a number of different strategies available for reducing the number of poles in the right-of-way, including the following:

� Increase the spacing between poles.� Use the same poles to carry multiple utilities (e.g., to carry both telephone, electric lines, and

luminaries).� Place poles on only one side of the street instead of both sides.� Selectively remove or relocate a limited number of poles from hazardous locations (i.e.,

intersections and horizontal curves).

A practical limitation to this strategy is that reducing the number of poles will likely requirelarger, more rigid poles to support more or heavier utility lines. This can be costly and the largerpoles could have an adverse effect on crash severity if a vehicle should strike a pole.

Figure B-3 can be used to estimate the reduction in utility pole crash frequency as a result ofreducing the number of poles (or pole density) in the right-of-way. The amount of the reductioncan be found by entering the nomograph with the two different pole densities and given poleoffset.

0.0

0.5

1.0

1.5

2.0

2.5

0 5 10 15 20 25 30

Average Pole Offset (ft)

Acc

iden

t F

requ

ency

(a

ccid

ents

/mile

/yea

r)

Pole Density Legend Low = 0-30 poles/mileModerate = 31-50 poles/mileHigh =>50 poles/mile

High

ModerateLow

Figure B-3. Relationship Between Frequency of UtilityPole Crashes and Pole Offset for Three Levels

of Pole Density (9).


B-12

Installing Breakaway Utility Poles

Using a breakaway design should be considered where poles have to be placed in vulnerablelocations that cannot economically be removed or related. Examples of where breakaway polesmight be effective include the following:

• gore areas,• the outside of sharp curves, and• opposite the intersecting roadway at T-intersections.

Details of both the breakaway utility pole and guy-wire connection designs are contained inFederal Highway Administration Report No. FHWA/RD-86/154, Safer Timber Utility Poles(10).

Placing Barriers Around Utility Poles

In those locations where it is not feasible to relocate and reduce the number of utility poles,shielding selected poles with guardrails or crash cushions may be warranted (of particular noteare the massive supports used for major electrical transmission lines within the clear zone or inother vulnerable locations). The reader should use the AASHTO Roadside Design Guide formore information on the types and warrants for installing roadside barriers.

Table B-6. Reduction in Utility Pole Crashes Due to Pole Relocation for Roadway Sections.

Pole OffsetBefore

Relocation, ft (m)

Expected Percent Reduction in Utility Pole Crashes (%)

Pole Offset After Relocation, m (ft)

6(1.8)

7(2.1)

8(2.4)

9(2.7)

10(3.1)

11(3.4)

12(3.7)

13(4.0)

14(4.3)

15(4.6)

20-30(6.1-9.2)

4 (1.2) 30 42 49 55 60 63 69 70 72 73 77

5 (1.5) 36 43 50 56 59 65 67 69 70 74

6 (1.8) 27 36 43 48 55 57 60 62 67

7 (2.1) 22 31 37 46 48 52 54 59

8 (2.4) 22 29 39 42 45 48 55

9 (2.7) 18 30 33 37 40 48

10 (3.1) 22 25 30 33 42

11 (3.4) 18 24 27 36

12 (3.7) 11 15 25

13 (4.0) 11 22

14 (4.3) 17


B-13

REFERENCES

1. Perchonok, K., T. Ranney, S. Baum, D. Morris, and J. Eppich. Hazardous Effects ofHighway Features and Roadside Objects, Volume I: Literature Review and Methodologyand Volume II: Findings, Report No. FHWA-RD-78-201 & 202, Federal HighwayAdministration, Washington, D.C. (September 1978).

2. Zegeer, C. V., J. Hummer, D. Reinfurt, L. Herf, and W. Hunter. Safety Effects of Cross-Section Design for Two-Lane Roads, -- Volumes I and II, Federal Highway Administration,Report No. FHWA/RD-87/008, Washington, D.C. (1987).

3. AASHTO, Roadside Design Guide, American Association of State Highway andTransportation Officials, Washington, D.C. (1996).

4. Zegeer, C. V., R. Stewart, D. Reinfurt, F. M. Council, T. Newman, E. Hamilton, T. Miller,and W. Hunter. Cost-Effective Geometric Improvements for Safety Upgrading ofHorizontal Curves, Report No. FHWA-RD-90-021, Federal Highway Administration,Washington, D.C. (1990).

5. Ross, H. E., K. K. Mak, and R. K. Lavingia. Safety of Driveways in Close Proximity toEach Other. TX-99/2946-S. (1998).

6. AASHTO. A Guide for Erecting Mailboxes on Highways. American Association of StateHighway Transportation Officials, Washington, D.C. (May 1984).

7. Ross, H. E., K. C. Walker, and W. J. Lindsay. The Rural Mailbox - A Little Known Hazard,Texas Transportation Institute, College Station, TX (1980).

8. Russell, E. R., M. J. Rys, and L. R. Contreras. “Guidelines for Roadside Safety on RuralLow-Volume Roads.” 1996 Compendium of Technical Papers, Institute of TransportationEngineers 66th Annual Meeting. MN. (1996).

9. Zeeger, C. V., and M. R. Parker. Cost-Effectiveness of Countermeasures for Utility PoleAccidents: User Manual, Report No. FHWA-IP-84-13, Federal Highway Administration,Washington, D.C. (July 1984).

10. Safer Timber Utility Poles. FHWA/RD-86/154. Federal Highway Administration,Washington, D.C. (1986).

APPENDIX C: ROADWAY CROSS SECTION

C-1

Roadway cross section includes the width of the travel way, the width and type of each lane, thewidth and type of shoulders, the cross slope of the pavement, and the slope of the side slopes. The American Association of State Highway and Transportation Officials (AASHTO) has setgeometric values for roadway features by functional classification. These values are presented inA Policy on Geometric Design of Highways and Streets, commonly known as the Green Book(1). The following is a summary of research findings presented in NCHRP Report 440 (AccidentMitigation Guide for Congested Rural Two-Lane Highways) (2) and from other sources onwidening the lane or shoulder, passing improvements, and two-way left-turn lanes (TWLTL).

WIDEN LANE OR SHOULDER

Travel lanes are the portion of the highway intended for use by through traffic. The lane width ofa two-lane road is measured from the centerline of the highway to the edgeline, or to theboundary between the travel lanes and the shoulder. Shoulders are the portion of the highwayimmediately adjacent to, and outside of, the travel lanes. Shoulders are typically designed andintended to accommodate occasional use by vehicles but not continual travel. Part or all of theshoulder may be paved. The lane and shoulder widths plus the median width comprise the totalroadway width. Total roadway width is among the most important cross-section considerationsin the safety performance of a two-lane highway. Generally, wider lanes and/or shoulders willresult in fewer crashes.

Numerous studies have been conducted in recent years to determine the effects of lane width,shoulder width, and shoulder type on crash experience; however, few of them were able tocontrol for roadside condition (e.g., clear zone, side slope), roadway alignment, and other factorswhich, together with lane and shoulder width, influence crash experience. Because lane andshoulder width logically affect some crash types (e.g., run-off-road, head-on) but not necessarilyother crash types (e.g., angle, rear-end), there is a need to express crash effects as a function ofthose related crash types.

Those crash types that research has shown that can be affected directly by lane and shoulderwidth improvements include the following:

� head-on, � run-off-road/fixed object,� run-off-road/rollover,� same direction sideswipes, and� opposite direction sideswipes.

A 1987 Federal Highway Administration study quantified the effects of lane width, shoulderwidth, and shoulder type on highway crash experience based on an analysis of data for nearly5000 mi (8050 km) of two-lane highway from seven states (3). The study controlled for manyroadway and traffic features, including roadside hazard, terrain, and average daily traffic. Crashtypes found to be related to lane and shoulder width, shoulder type, and roadside conditioninclude run-off-road (fixed object, rollover, and other run-off-road crashes), head-on, andopposite- and same-direction sideswipe crashes, which were termed as “related crashes.” If a


C-2

user knows only the number of total crashes on the section, Table C-1 gives factors to convertbetween total and related types. Since ADT and terrain are factors which influence theproportion of various crash types on a section, the table provides adjustments for these factors. The expected effects of lane and shoulder widening improvements on related crashes follow.

Table C-1. Factors to Convert Total Crashes to Related Crashes on Two-Lane Rural Roads (3).

ADT (vpd) Terrain Adjustment Factors

Flat Rolling Mountainous

5001000200040007000

10000

.58

.51

.45

.38

.33

.30

.66

.63

.57

.48

.40

.33

.77

.75

.72

.61

.50

.40

Note: Related crashes include run-off-road, head-on, opposite-direction, and same-direction sideswipe.

Table C-2 summarizes the percent reduction in crash frequency as a result of increasing lanewidths. Significant reduction in crash frequency can be achieved with only minor increases inlane widths. For example, widening a lane by as little as 1 ft (0.3 m) (e.g., from 10- to 11-ft [3.1to 3.4 m] lanes) can reduce the frequency of related crashes by as much as 12 percent. Wideninga lane by 4 ft (1.2 m) (e.g., from 8- to 12-ft [2.4 to 3.7 m] lanes) could result in a 40 percentreduction in related crash types.

It should be noted, however, that increasing lane widths above a total of 12 to 15 ft (3.7 to 4.6 m)has little benefit in reducing crash frequency. In fact, when lane widths become too wide, driverscan become confused as to the total number of lanes on a roadway. This can lead to an increasein some types of crashes, especially same-direction sideswipes.

Table C-2. Percentage of Crash Reduction of Related Crash Types for Lane Widening Only (3).

Amount of LaneWidening, m (ft)

Percent Reduction inCrash Types (%)

0.3 (1)0.6 (2)0.9 (3)1.2 (4)

12233240


An expert panel recently convened as part of an FHWA study confirmed the Zegeer et al. studyas the most reliable assessment of the effect of lane width on safety for two-lane highways with


C-3

ADTs over 2000 veh/day (3, 4). Table C-3 illustrates the recommendations of that expert panelexpressed as crash modification factors (or relative crash frequencies). For example, the crashmodification factor of 1.50 for 9-ft (2.8 m) lanes in Table C-3 implies that a two-lane highwaywith 9-ft (2.8 m) lanes would be expected to experience 50 percent more crashes of the typespecified than a two-lane highway with 12-ft (3.7 m) lanes. In using Table C-3, it should beassumed that the safety performance of lane width less than 9 ft (2.7 m) is the same as that shownfor 9-ft (2.8 m) lanes, and that the safety performance for lanes over 12 ft (3.7 m) wide is thesame as that shown for 12-ft (3.7 m) lanes. Interpolation between the values shown in Table C-3is encouraged.

Research results concerning reductions in related crashes due to widening paved or unpavedshoulders are listed in Table C-4. For example, widening 2-ft (0.6 m) gravel shoulders to 8 ft (2 m) will reduce related crashes by 35 percent (i.e., for a 6-ft (1.8 m) increase in unpavedshoulders). Adding 8-ft (2.4 m) paved shoulders to a road with no shoulders will reduceapproximately 49 percent of the related crashes (3). It should be noted that the predicted crashreductions given in Tables C-3 and C-4 are valid only when the roadside characteristics (sideslope and clear zone) are reestablished as before the lane or shoulder widening.

Table C-3. Crash Modification Factors for Lane Width on Rural Two-Lane Highways (3).

Lane Width, ft (m) Crash Modification Factor a

9 (2.8) 10 (3.1)11 (3.4)12 (3.7)

1.501.301.151.00

aRelative crash frequency for run-off-road, head-on, and opposite-direction sideswipe crashes.

Table C-4. Percentage of Crash Reduction of Related Crash Types for Shoulder Widening Only (3).Shoulder Widening

per Side, ft (m)Percent Reduction in Related

Crash Types (%)

Paved Unpaved

(2) 0.6(4) 1.2(6) 1.8(8) 2.4

16294049

13 253543


The expert panel discussed above also found the Zegeer et al. study to be the most reliableassessment of the effect of shoulder width on safety for two-lane highways with ADTs over 2000veh/day (3, 4). Table C-5 illustrates the recommendations of that panel for the effect of shoulder


C-4

width on safety expressed as crash modification factors. Table C-6 shows similar results for theeffect of shoulder type. Interpolation within these tables is encouraged, but extrapolation beyondtheir limits is not.

The crash modification factors for lane width, shoulder width, and shoulder type can becombined by multiplying them together. For example, the crash modification factor for acombination of 11-ft (3.4 m) lanes and 4-ft (1.2 m) gravel shoulders can be determined as:

(1.15)(1.15)(1.01) = 1.34

This implies that a two-lane highway with 11-ft (3.4 m) lanes and 4-ft (1.2 m) gravel shoulderswould experience 34 percent more related crashes than a two-lane highway with the nominalcondition of 12-ft (3.7 m) lanes and 6-ft (1.8 m) paved shoulders.

Table C-5. Crash Modification Factors for Shoulder Width on Rural Two-Lane Highways (4).

Shoulder Width, (ft) m Crash Modification Factor a

(0) 0(2) 0.6(4) 1.2(6) 1.8(8) 2.4

1.501.301.151.000.87

a Relative crash frequency for run-off-road, head-on, andopposite-direction sideswipe crashes.

Table C-6. Crash Modification Factors for Shoulder Types on Two-Lane Highways (4).

Shoulder Type

Crash Modification Factor a

Shoulder Width, ft (m)

0 (0) 1 (0.3) 2 (0.6) 3 (0.9) 4 (1.2) 6 (1.8) 8 (2.4) 10 (3.1)

PavedGravel

CompositeTurf

1.001.001.001.00

1.001.001.011.01

1.001.001.021.03

1.001.011.021.04

1.001.011.031.05

1.001.021.041.08

1.001.021.061.11

1.001.031.071.14

a Relative crash frequencies for run-off-road, head-on, and opposite-direction sideswipe crashes.

Other studies have also examined the benefits of widening pavements. Table C-7 providespercent reductions in total, single-vehicle, and head-on crashes due to widening pavements oradding full-width paved shoulders. Although sample sizes are small in certain cells, these resultssupport the findings in other studies in terms of the beneficial effects of lane and shoulderwidening, the types of crashes reduced, and the relative magnitude of the effects of widening.


C-5

Table C-7. Summary of Crash Reductions for Pavement Widening Projects (5, 6).

Type of ProjectExpected Percent Reduction in Crashes

ADT Range (vpd) Total CrashesSingle-Vehicle

CrashesHead-OnCrashes

Widening 20- to 24-ft(6.1 to 7.3 m) pavementto 28 ft (8.5 m)

0-3000 16.0 (C) 22.0 (C) 45.0 (C)


<5000 35.0 (C) (s) 49.0 (C) (s) 48.0 (C) (s)


>5000 29.0 (C) (s) 22.0 (C) (s) 51.0 (C) (s)

Adding full-widthpaved shoulders totwo-lane roads

1000-30003000-50005000-7000

27.0 (T) (s)12.5 (T)

17.6 (T) (s)

55.0 (T) (s)21.4 (T) (s)

0.0 (T)

UnknownUnknownUnknown

Notes:(C) = values from the Rinde study (5) in California(T) = values from the Turner et al. (6) study in Texas(s) = significant at the 95 percent level of confidence for (C) sites and 90 percent confidence level for the (T)

sites.The single-vehicle and head-on crash percentages for California were adjusted by 4 to 6 percent to account forexternal effects and are now on the same basis as total crashes. These values are only for two-lane rural roads.

A 1987 study for the Texas Department of Transportation investigated the relationship betweencrash rate and crown width (surface width) on rural, two-lane, farm-to-market roads (7). Thepercent reduction factors determined for single-vehicle crashes are listed in Table C-8. Thereduction factors were estimated based upon regression equations of approximately 1400 mi(2254 km) of roadways and 4000 crashes. The analysis indicated that widening existing rural,two-lane, farm-to-market roads carrying over 1000 vehicles per day to a minimum of 22, 24, or26 ft (6.7, 7.3, or 7.9 m) would yield benefit cost ratios of 1.07, 1.14, and 1.17, respectively. Themajor findings from the study were:

• Surface width has no demonstrable effect on multi-vehicle crash rate on rural, two-lane,farm-to-market roads with AADTs up to 1500.

• Surface widening can reduce single-vehicle crash rate on rural, two-lane, farm-to-marketroads with AADTs up to 1500.

• While surface widening can reduce single-vehicle crash rate on rural, two-lane, farm-to-market roads with AADTs in excess of 400, such action is not warranted (i.e., not costbeneficial) at AADTs below 1000.


C-6

Table C-8. Single-Vehicle Crash Reduction Factors (%) Associated with Surface Widening in Three ADT Categories (7).

AADT ExistingSurface Width,

ft (m)

Resurfaced Width, ft (m)

20 (6.1) 22 (6.7) 24 (7.3) 26 (7.9)

401 to 700 18 (5.5)20 (6.1)22 (6.7)24 (7.3)

7 137

19137

2519137

701 to 1000 18 (5.5)20 (6.1)22 (6.7)24 (7.3)

12 2312

322313

41332413

1001 to 1500 18 (5.5)20 (6.1)22 (6.7)24 (7.3)

14 2715

382816

49403017

PASSING IMPROVEMENTS

A majority of two-lane highways carry relatively low-traffic volumes and experience fewoperational problems; however, some higher volume two-lane highways experience safety andoperational problems. Often such problems can be related to inadequate geometry (steep grades,poor sight distance) and the lack of passing opportunities (due to heavy oncoming traffic and/orpoor sight distance). While a major reconstruction project may be used to reduce the problem(e.g., widening to a four-lane facility or major alignment changes), other lower cost alternativeshave been used successfully to reduce crash operational problems.

The following are strategies for adding passing opportunities to a basic two-lane highway toimprove operations and safety:

� passing lanes,� climbing lanes,� short four-lane sections,� turnouts, and � shoulder use sections (i.e., shoulders are used as driving lanes).

These countermeasures are illustrated in Figure C-1.

The need for passing opportunities on a two-lane road arises when the demand for passingopportunities exceeds their supply. The supply of passing opportunities on a two-lane roaddepends on the availability of passing sight distance and gaps in the opposing traffic stream. When the demand exceeds the supply, traffic platoons develop and grow as faster vehicles catch


C-7

up with slower ones and are unable to pass. Passing lanes and use of shoulders and turnouts byslow vehicles can increase the opportunity for the faster vehicles to pass the slower vehicles.

A passing lane is an added lane provided in one or both directions of travel on a conventionaltwo-lane highway to improve passing opportunities. This definition includes passing lanes inlevel or rolling terrain, climbing lanes on grades, and short four lane sections. The length of theadded lane can vary from 1000 ft (305 m) to as much as 3 mi (5 km). When passing lanes areprovided at an isolated location, their function is generally to reduce delays at a specificbottleneck, and the location of the passing lane is dictated by the needs of the specific trafficoperational problem encountered. When passing lanes are provided to improve overall trafficoperations over a length of road, they are often constructed systematically at regular intervals (8).

Figure C-1. Typical Operational Treatments Used on Two-Lane Highways (8).

A turnout is a widened, unobstructed shoulder area on a two-lane highway that allows slow-moving vehicles to pull out of the through lane to permit following vehicles to pass. Turnouts


C-8

are relatively short, generally less than 600 ft (190 m) in length. They have been used mostwidely in the western United States; however, they are applicable to any winding or mountainoustwo-lane highway with limited passing opportunities (8).

The primary purpose of the shoulder on a two-lane highway is to provide a stopping and recoveryarea for disabled or errant vehicles. In some parts of the United States there is a long-standingcustom where adequate paved shoulders are provided for slow-moving vehicles to move to theshoulder when another vehicle approaches from the rear and return to the travel lane after thevehicle has passed. Shoulder driving can provide great operational flexibility for two-lanehighways; however, this option may not be appropriate at all locations because driving on theshoulder is illegal in most states.

An approach for allowing the use of paved shoulders by slow-moving vehicles is to permit thepractice at selected sites designated by specific signing. Signs are placed at both the beginningand end of the highway section where shoulder use is allowed. This approach results in ahighway agency encouraging shoulder use by slow-moving vehicles only where it has beenestablished that additional passing opportunities are needed and where the shoulder is structurallyadequate to handle the anticipated traffic loads. In some cases, the use of the shoulder isrestricted to daylight hours only.

Table C-9 summarizes the results of a research study that examined how sections of highwaywhere the above mentioned countermeasures were implemented compared to adjacent“untreated” two-lane highway sections. Reductions in crash frequencies of 25 to 40 percent werereported for passing lanes, short four-lane sections, and turnout lanes (8, 9). Note that thesereductions are based on sites that carried predominantly higher traffic volumes than average two-lane sections. Thus, the reductions shown in Table C-9 may not apply to low-volume two-laneroads.

The reader should use caution regarding the crash effects of these design alternatives becausecrash experience may vary widely depending on the specific traffic and site characteristics. Inaddition, not all of these alternatives are appropriate for all possible roadway sections. Also,while such alternatives may reduce some safety and operational problems, other problems may becreated in some cases. More detailed guidelines for optimal use of these design alternatives aregiven in an Informational Guide by Harwood and Hoban (8).

TWO-WAY LEFT-TURN LANE

Two-way left-turn lanes (TWLTLs) are paved areas in the highway median marked to provide adeceleration and storage area for vehicles traveling in either direction to make left turns intointersections and driveways. TWLTLs have been used for many years on urban and suburbanarterial streets with commercial development to improve safety and to reduce delays to throughvehicles caused by turning traffic. Highway agencies have recently begun to use TWLTLs inrural and urban fringe areas to obtain these same types of operational and safety benefits.


C-9

Table C-9. Crash Reductions Related to Five Design Alternatives, as Compared to a BasicTwo-Lane Road Design (8, 9).

Design Alternative Type of AreaPercent Reduction in Crashes a

Total Crashes F + I Crashes b

Passing lanesShort four-lane sectionTurnouts Shoulder use section

RuralRuralRuralRural

253530(c)

304040(c)

Notes:a These values are only for two-lane roads in rural or suburban areas.b F + I = fatal plus injury crashesc no known significant effect

TWLTLs are particularly appropriate at locations where high left-turn volumes are distributedacross a range of driveways or intersections and at locations where there is a documented patternof left-turn crashes spread over several intersections or driveways. Care should be taken not tooveruse TWLTLs on two-lane highways because passing is prohibited in TWLTL sections. Ifused in areas with minimal development, TWLTLs can be operationally detrimental by denyingdrivers the opportunity to pass slow-moving vehicles, without any corresponding safety benefit. When evaluating whether to install a TWLTL, highway agencies should consider the availabilityof passing opportunities on the adjacent highway section. If the only good passing zone for milesin either direction is replaced by a TWLTL, illegal passing maneuvers are likely, and thepotential for conflicts between passing and turning vehicles is increased.

TWLTLs are effective in reducing left-turn crash rates and rear-end crashes. TWLTLs have beenfound to reduce crash rates by approximately 35 percent when installed at urban and suburbansites, primarily on multilane highways (10). Comparable crash reduction effectiveness was foundby Harwood and St. John (9) for installation of TWLTLs on two-lane highways in urban fringeareas. In rural areas, the number of crashes at candidate TWLTLs on two-lane highways is small,but TWLTLs can reduce these crashes by up to 85 percent.

A field study of traffic conflicts and erratic maneuvers at four rural TWLTL sites on two-lanehighways found only one problem that was consistent: illegal passing in the TWLTL wasobserved by a relatively small fraction (0.4 percent) of vehicles (9). Since it is evident that somedrivers will pass illegally in TWLTLs, a careful evaluation of any proposed TWLTL installationthat would eliminate an existing passing zone is recommended.


C-10

REFERENCES

1. AASHTO. A Policy of Geometric Design of Highways and Streets, American Associationof State Highway Transportation Officials, Washington, D.C. (2001).

2. Fitzpatrick, K., D. W. Harwood, I. B. Anderson, and K. Balke. Accident Mitigation Guidefor Congested Rural Two-Lane Highways. NCHRP Report 440. Transportation ResearchBoard, Washington, D.C. (2000).

3. Zegeer, C. V., J. Hummer, D. Reinfurt, L. Herf, and W. Hunter. Safety Effects of Cross-Section Design for Two-Lane Roads, – Volumes I and II, Federal Highway Administration,Report No. FHWA/RD-87/008, Washington, D.C. (1987).

4. Harwood, D. W., F. M. Council, E. Hauer, W. E. Hughes, and A. Vogt. Prediction of theExpected Safety Performance of Rural Two-Lane Highways, Report No. FHWA/RD-4584-09, Federal Highway Administration, McLean, VA (November 1998).

5. Rinde, E. Accident Rates vs. Shoulder Widths, Report No. CA-DOT-RT-3147-1-77-01,Federal Highway Administration, California DOT, Sacramento, CA (1977).

6. Turner, D. S., D. B. Fambro, and R. O. Rogness. “Effects of Paved Shoulders on AccidentRates for Rural Texas Highways,” Transportation Research Record No. 819.Transportation Research Board, Washington, D.C. (1981).

7. Griffin, L. I., and K. K. Mak. The Benefits To Be Achieved from Widening Rural, Two-Lane, Farm-to-Market Roads in Texas. Texas Transportation Institute, College Station, TX(June 1987).

8. Harwood, D.W., and C. J. Hoban. Low-Cost Methods for Improving Traffic Operations onTwo-Lane Roads: Information Guide, Report No. FHWA-IP-87-2, Federal HighwayAdministration, Washington, D.C. (January 1987).

9. Harwood D. W., and A. D. St. John. Passing Lanes and Other Operational Improvementson Two-Lane Highways, Report No. FHWA/RD-85/028, Federal Highway Administration,Washington, D.C. (December 1985).

10. Harwood, D. W., and J. C. Glennon. “Selection of Median Treatment For Existing ArterialHighways,” Transportation Research Record 681, Transportation Research Board,Washington, D.C. (1978).

APPENDIX D: ALIGNMENT

D-1

Roadway alignment includes straight sections, horizontal curves, roadway grades, and verticalcurves. The American Association of State Highway and Transportation Officials (AASHTO)has set geometric values for roadway features by functional classification. These values arepresented in A Policy on Geometric Design of Highways and Streets, commonly known as theGreen Book (1). Traffic control devices, such as signs and markings, can also assist in creating aroadway that performs well and safely. Information on these devices is presented in the Manualon Uniform Traffic Control Devices (MUTCD) (2).

Research has been conducted to evaluate the safety of various alignments. In some cases, theseefforts have identified definitive relationships between design values and safety. The findingshave allowed the development of crash reduction estimates that would be expected due to relatedroadway safety improvements. In other cases, the evidence only provides an estimate orsuggestions of how to improve a roadway. The findings from these different studies provideresults that can be used to improve the design of rural two-lane highways.

HORIZONTAL ALIGNMENT

Crash studies indicate that horizontal curves experience a higher crash rate than tangents, withrates ranging from one and a half to four times greater than tangent sections (1, 2, 3). Pastresearch has identified a number of traffic, roadway, and geometric features that are related to thesafety of horizontal curves. These factors include the following (3):

� traffic volume on the curve and traffic mix (e.g., percent trucks);� curve features (degree of curve, length of curve, central angle, superelevation, presence of

spiral, or other transition curves);� cross-sectional curve elements (lane width, shoulder width, shoulder type, and shoulder

slope);� roadside hazard on the curve (clear zone, side slope, rigidity, and types of obstacles);� stopping sight distance on curve (or on curve approach);� vertical alignment on horizontal curve;� distance to adjacent curves;� presence/distance from curve to the nearest intersection, driveway, bridge, etc.;� pavement friction;� presence and type of traffic control devices (signs and delineation); and� others.

Previous studies show clearly that sharper curves are associated with higher crash rates thanmilder ones (4, 5, 6, 7). The types of crashes generally found to be more represented on curvescompared to tangents included more severe (fatal and A-type injury) crashes, head-on andopposite-direction sideswipe crashes, fixed-object and rollover crashes, crashes at night, andcrashes involving drinking drivers. Based on a larger sample of 10,900 horizontal curves inWashington State, the distribution of curve crashes by severity and type were determined asshown in Table D-1.


D-2

Of all the factors that affect the design of horizontal curves, the degree of curvature is the bestpredicted of the crash potential (4, 8). Simply stated, flatter curves (i.e., curves with low degreesof curvature) are more likely to have fewer crashes than sharper curves.

Geometric countermeasures that are used to improve safety along horizontal curves include thefollowing:

• curve flattening,• roadway widening on curves,• superelevation improvements, and• roadside improvements on curves.

Traffic control devices used at horizontal curves include delineation treatments and curvewarning signs. These treatments are discussed in the following traffic control devices section. Information on roadside improvements is provided in Appendix B.

Table D-1. Summary of Crash Statistics on Washington State Curve Sample (5).Variable Frequency Percentage

Total crashes 12,123 100.0

PDO crashesInjury crashesFatal crashes

65005359264

53.644.22.2

People injuredPeople killed

8434*314*

N/AN/A

Head-on crashesOpposite-direction sideswipe crashesFixed-object crashesRollover crashesSame-direction sideswipeRear-end both movingOther collision types

517468504518741393033777

4.33.941.615.51.12.531.2

Dry-road crashesWet-road crashesSnowy/icy road crashes

691426092600

57.021.521.4

Daylight crashesDark, dawn, dusk crashes

68285295

56.343.7

* These are numbers of people injured or killed, and not the number of crashes in which someone was injuredor killed.

Curve Flattening

Curve flattening refers to reconstructing a horizontal curve to make it less sharp (i.e., a largerradius value or a lower degree of curve). The IHSDM (Interactive Highway Safety DesignModel) crash prediction model for two-lane highways uses the following equation to determinethe expected crash frequency of a horizontal curve relative to a tangent roadway:


D-3

A M F = 1 55

80 20 0125

1 55

. .

.

.

LR

L

c

c

+ −

Where: AMF = crash modification factor Lc = length of horizontal curve (mi)

R = radius of curvature (ft)S = 1 if spiral transition curve is present or 0 if spiral transition curve is not

present

This equation is based on the result of Zegeer et al. (5). In applying this equation to a curve witha spiral transition, Lc represents the length of the circular portion of the curve. When a curve isflattened (i.e., when its radius is reduced) but its central angle remains the same, the length of thecurve increases. This must be considered when evaluating the expected effect of curve-flatteningprojects. The expected crash rate of the longer curve with larger radius should be compared tothe shorter, sharper curve plus two tangent sections on either end.

Roadway Widening on Curves

Wider lanes and shoulders on curves are also associated with a reduction in curve-relatedcrashes. Percent reductions in total crashes are given in Table D-2 for improvements involvingwidening lanes and/or shoulders on horizontal curves (5). From the left column of the table, theuser should select the amount of lane or shoulder widening that is proposed for the project.

Table D-2. Percent Reduction in Crashes Due to Lane and Shoulder Widening onHorizontal Curves (5).

Total Amount of Lane or Shoulder Widening, ft (m)

Percent Crash Reductions (%)

Total Per Side Lane Widening Paved Shoulder Widening

Unpaved ShoulderWidening

2 (0.6)4 (1.2)6 (1.8)8 (2.4)10 (3.1)12 (3.7)14 (4.3)16 (4.9)18 (5.5)20 (6.1)

1 (0.3)2 (0.6)3 (0.9)4 (1.2)5 (1.5)6 (1.8)7 (2.1)8 (2.4)9 (2.7)

10 (3.1)

5121721——————

481215192125283133

371013161821242629

The columns in Table D-2 provide the expected percent reduction in total crashes for wideninglanes, paved shoulders, and unpaved shoulders, respectively. For example, assume a 20-ft


D-4

(6.1 m) roadway (i.e., two 10-ft (3.1 m) lanes with no shoulder) is to be widened to 22 ft (6.7 m)of paved surface with 8-ft (2.4 m) gravel shoulders (i.e., 16 ft (4.9 m) total of shoulder widening). From Table D-2, these improvements would reduce curve crashes by 5 percent (due to lanewidening) and 24 percent due to widening unpaved shoulders by 2.4 m (8 ft). Note that the 5percent and 24 percent crash reduction values cannot merely be added numerically. The properprocedure for combining two or more crash reduction factors is discussed in Zegeer et al. (5).

Superelevation Improvements

Superelevation is the amount of “banking” or cross-slope of a horizontal curve. A number ofstudies have attempted to link superelevation to crash causation. One study by Zador et al. noteddeficiencies in available superelevation at fatal crash sites compared with nearby control sites (9). In the 1991 FHWA study, a small but significant crash effect of too little superelevation wasnoted (5). The authors concluded that curve sites with a superelevation “deficiency” havesignificantly worse crash experience than curves with a proper amount of superelevation. Thesuperelevation deficiency, eD, is defined as the difference between the recommendedsuperelevation according to the Green Book (1), (eR), and actual superelevation (eA) or eD = eR - eA.

Table D-3 shows the percent reduction in total curve crashes due to improving superelevation. To illustrate the use of the table, assume the actual superelevation (eA) on a curve is 0.04 and theAASHTO recommended superelevation (eR) for a particular curve design is 0.06. Thiscorresponds to a superelevation deficiency eD of 0.02. According to Table D-3, which is amodification of the Zegeer et al. (3) results by the expert panel that developed the IHSDM crashprediction algorithm (10), a horizontal curve with a superelevation deficiency of 0.02 wouldexperience 6 percent more crashes than a similar horizontal curve with no superelevationdeficiency.

Table D-3. Crash Modification Factor for Superelevation Deficiency on Two-Lane Highway Horizontal Curves (10).

Superelevation Deficiency Crash Modification Factor

0.000.010.020.030.04

1.001.001.061.091.12

It should be noted that the 1991 study also investigated the safety effect of too muchsuperelevation. No adverse effects were found based on available data. Current design policy isimplemented with an assumed upper limit on superelevation for areas with snow and ice. Thepresumption is that excess superelevation produces sliding down the curve under low-speedconditions and hence increases crash potential. While this condition could theoretically occur at


D-5

low-speed curve locations with sharp curvature and a high rate of superelevation, no evidencewas found of any such significant adverse safety effects.

VERTICAL ALIGNMENT

The vertical alignment selected for a highway is a compromise between existing terrain, safety,and construction cost. It is described by both vertical lines or grades and vertical curvesincluding the sags and crests. The design of crest vertical curves is influenced by the differencebetween the grades and the stopping distance selected for the roadway. Stopping sight distance isthe sight distance available on a roadway that would permit a below-average operator or vehicletraveling at or near the design speed of the roadway to stop before reaching a stationary object inits path. For crest vertical curves, the sight distances are determined for drivers to see over thetop of the hill to objects on the other side. For sag vertical curves, the sight distances aredetermined for drivers seeing at night from the vehicles’ headlights.

Table D-4 shows that downgrade crashes are more frequent and result in higher percentages ofinjuries and fatalities than upgrade crashes. Also, injury and fatality rates on vertical curves arehigher than on level or upgrade locations. The crash rate for downgrades is 63 percent higherthan for upgrades, assuming that upgrades have as much vehicular traffic as downgrades (11).

Table D-4. Crash Frequency and Severity by Vertical Alignment (11).

VerticalAlignment

Number ofCrashes

Percent of TotalCrashes

Percent Injured Percent Fatal

LevelUpgradeDowngradeUp on crestDown on crest Up on sagDown on sag

2001943

1533373461258211

34.616.326.56.58.04.53.7

53.655.658.459.562.657.861.7

4.73.95.16.05.96.36.8

Total KnownUnknownTotal

578021927972

100.0

The elements of vertical alignment that are believed to influence safety include the steepness andlength of grade and the vertical curve design. Geometric countermeasures for vertical alignmentinclude minimizing the effects of slower moving vehicles by providing opportunities to pass andincreasing sight distance on or minimizing hazards within a vertical curve. Information onproviding passing opportunities is presented in Appendix C. Information on grade and on sightdistance on crest vertical curves follows.

A recent study by Miaou developed a crash modification factor for the effect of vertical grade oncrash frequency on rural two-lane highways (12). This factor, shown in Table D-5, is equivalent


D-6

to an increase in crash frequency of 1.6 percent per percent grade. The crash modification factorsshown in Table D-5 have been incorporated in the IHSDM crash prediction algorithm (10).

Table D-5. Crash Modification Factors for Grade of Roadway Sections (10).

Grade (%)

0 2 4 6 8

1.00 1.03 1.07 1.10 1.14

Note: This factor can be expressed as an effect of 1.6 % per percent grade.

Recent studies on vertical curves found:

• Crash rates on rural two-lane highways with limited stopping sight distance are similar to thecrash rates on all two-lane rural highways (13).

• Vertical curves with stopping distances less than 311.5 ft (95 m) had more crashes thanvertical curves with very long stopping sight distances. The largest increase in crashesoccurred at the study sites that had the shortest stopping sight distances (14).

• Stopping sight distances ranging from 328 to 426.5 ft (100 to 130 m) did not affect crashrates unless an intersection was within the limited sight distance section (15).

Thus, for the range of conditions studied, limited stopping sight distance does not appear to causea safety problem. The following recommendations regarding the safety effects of limitedstopping distance were made (13):

• Many design criteria are based on parameters associated with the interaction between drivers,vehicles, and roadways. The resultant design criteria should be greater than or equal to theminimum requirements for safety.

• Based on the literature and on this study, the minimum stopping sight distance for safeoperations and a 56 mph (90 km/h) speed is somewhere between 311.7 to 361 ft (95 to 110m). The values are less than the minimum design values in the 1994 AASHTO Green Book;therefore, the AASHTO stopping sight distance represents acceptable values for design. Thethreshold for safe operations may increase when hazards, such as intersections or horizontalcurves, are located within the limited stopping sight distance section.

• Because there are no apparent safety benefits from providing stopping sight distances longerthan 360 ft (110 m) (when other hazards are not present), improvements other thanlengthening a limited stopping sight distance crest vertical curve may provide a moreeffective use of available funds.

TRAFFIC CONTROL DEVICES

An informed driver with sufficient time to respond to a situation can avoid making seriousdriving errors. Conversely, inadequate information or time to respond to a situation results in ahigh probability of an erratic maneuver and high potential for a crash. Thus, communicating


D-7

clear and concise information that is timely and meaningful to drivers is essential to a safedriving environment. Signs, pavement markings, and delineators can provide drivers withadditional information concerning the roadway such as unexpected or atypical situations.Delineation refers to any method of defining the roadway operating area for the driver (16). Delineation has been defined as one or more devices that regulate, warn, or provide trackinginformation and guidance to the driver. These devices include the following delineationmaterials: painted markings, thermoplastic and other durable markings, raised pavement markers,and post-mounted delineators. Warning signs are also considered part of the delineation systemwhen used to complement standard delineation in special areas, such as horizontal curves. Forthis document, delineation techniques have been divided into signs; pavement markings,including both paint and thermoplastic markings; and delineators, such as raised pavementmarkers and post-mounted delineators.

The Roadway Delineation Practices Handbook (16) was developed to assist in making decisionsabout roadway delineation systems. The Handbook supplements the MUTCD (2) by offeringimplementation guidelines. The contents cover current and newly developed devices, materials,and installation equipment, and presenting each item’s expected performance based on actualexperience or field and laboratory tests. Individual chapters cover the characteristics ofretroreflection and quality assurance, driver visibility needs, traffic points, preformed tapers,raised pavement markings and other marking materials, post-mounted delineators and otherdelineation devices, and administrative and management issues and practices.

Signs

A recent TxDOT study has evaluated guide signing for rural highways with the final task of thestudy devoted to the development of a field book for guide signing on conventional highways(17) . The Sign Crew Field Book is intended to provide field sign personnel with informationbeyond that contained in the Texas MUTCD or the TxDOT Traffic Control Standard Sheets sothat guide signing can be applied in a more uniform manner.

A 1980s study in Ohio examined the effectiveness of advisory speed signs used in conjunctionwith curve warning signs (18). The results of the test-driver study indicated that drivers, on theaverage, look about two times at a warning sign (fixation duration 0.5 to 0.6 seconds). Basedupon the findings from the 40 test drivers, the author concluded that advisory speed signs are notmore effective in causing drivers to reduce their speeds through curves than curve and turn signsalone. Other studies have also found that various sign treatments for reducing traffic speeds inthe vicinity of horizontal curve have generally been ineffective (19, 20).

Pavement Markings

Pavement markings are used to provide regulations and warnings to the driver. They can be usedeither alone or to supplement the regulations or warnings of other devices such as traffic signs orsignals. Longitudinal pavement markings are used to indicate lane lines and edgelines. Yellowis used to separate traffic in opposing directions while white is used to separate traffic flowing inthe same direction or mark the right edge of the pavement. Whether the lines are solid or broken


D-8

and wide communicates restrictions along the roadway. Markings are more effective atcommunicating this type of information than signs. Right- and left-side edgelines arerecommended for all roadways with any substantial traffic volumes.

Pavement markings studies have examined the effectiveness of edgeline and centerline markingsand whether there are benefits to using wider markings in certain areas. The use of illusion-creating markings has also been investigated along with unique markings selected to reducespeeds prior to a horizontal curve.

Edgeline and Centerline Markings

The use of 4-in (10.2 cm) edgelines significantly reduced the number of crashes as comparedwith those sites with no edgelines (21). The use of 4-in (10.2 cm) edgelines has also shown asignificant reduction in the number of crashes at access points (i.e., driveways and intersections)(22). Adding edgelines and centerlines to roadways where no delineation has been providedreduced crashes by 36 percent in a 1970s study (23). Adding centerlines reduced crashes by 29percent; adding edgelines to centerlines yielded an 8 percent reduction. A Kansas study involvedcontrol and treatment sites comprising 384 mi (618.2 km) of rural highway servicing between550 and 3600 vehicles per day. Using these findings, it was determined that edgelines will yieldbenefits exceeding their costs if an average of one non-intersection crash occurs annually every15.5 mi (25 km) of roadway (24).

Wide Markings

Several states have experimented with using 8-in (20.3 cm) edgelines to prevent run-off-road(ROR) crashes (25, 26, 27). In general, the effectiveness of 8-in (20.3 cm) edgelines to reducerun-off-road crashes is questionable. Their use is recommended for rural roadways where thepavement width is at least 24 ft (7.3 m), the shoulders are unpaved, and the average day traffic(ADT) is between 2000 and 5000 vehicles per day. Eight-inch edgelines are not recommendedon two-lane, rural roads with the following exceptions:

• frequent heavy snowfall and use of deicing materials and abrasives that tends to deteriorateedgelines,

• pavement widths of less than or equal to 6.7 m (22 ft), and• roads having paved shoulders over 1.8 m (6 ft) wide.

Eight-inch edgelines may be appropriate as a safety improvement when applied at spot locationssuch as isolated horizontal curves and approaches to narrow bridges.

Unique Markings

Transverse pavement markings have been tested to determine if drivers will slow down inadvance to a curve. Average traffic speeds were reduced from 41.3 to 33.9 mph (66.5 to 54.6km/h) one week after markings were installed at one site and, six months after treatment, theaverage speed was 34.8 mph (56.0 km/h) — 16 percent less than observed during the baseline


D-9

period (20). Another study (28) also reported reductions in traffic speeds, most notably highspeeds, resulting from pavement markings designed to make the roadways appear narrower at thebeginning of the curves. The pavement markings shown in Figure D-1 were tested to determinewhether excessive traffic speeds at rural and suburban two-lane roadway locations with sharphorizontal curves could be reduced. The pavement markings were associated with a decrease invehicle speed of approximately 6 percent overall and 7 percent during daytime and late nightperiods (29).

Figure D-1. Pavement Marking and SpeedMeasurement Locations

for Retting and Farmer’s Study (29).

Griffin and Reinhardt (30) reviewed theavailable literature on two illusion-creating pavement marking patterns. Themarkings were developed and used in thelast 20 years to reduce traffic speeds andtraffic crashes that result from driverinattention and habituation to high-speeddriving. The marking patterns were theconverging chevron pavement markingpattern and the transverse bar pavementmarking pattern (most often used at theapproaches to traffic circles). Based on areview of 10 different studies of theeffects, the following was found:

� Most of the studies that werereviewed indicated that traffic speedscould be reduced by the application oftransverse bar markings.

� Some studies suggested that thespeed-reduction effectiveness of thesepatterns can be maintained for manymonths; others suggest the benefits ofthe markings are transitory and fadewithin a matter of days or weeks.

� When transverse bars were used inconjunction with pavementdiscontinuities (i.e., rumble strips),speed reduction was enhanced, butspeed variability tended to increase.

� Transverse bar markings may reducetraffic speeds because the patternsmay be functioning as a warningsignal rather than from the illusionthat drivers are traveling too fast.


D-10

Raised Pavement Markers

The application of raised pavement markers provides several benefits including increaseddelineation of the driving path of the roadway, increased ability to “track” the roadway, increasedreflectivity under wet-weather conditions, and increased tactile and auditory warning to driverswhen crossing the markers (31). Despite the clear advantages of RPMs, several studies haveindicated an increase in nighttime crashes when RPMs are present (32, 33, 34), perhaps as aresult of an increased sense of confidence in the driving task.

A study was performed to note driver responses to the application of raised pavement markers(RPM) by measuring changes in speed and encroachment distances into the opposing travel laneafter varying the spacing intervals of the markers (35). The RPM were spaced at a 40-ft (12 m)and 20-ft (6 m) spacing. The study used two rural minor arterial sites. The study recommendedat least a 40-ft (12 m) spacing interval for the markings. They found that spacings below thisvalue were shown to be no more effective in daylight conditions and are more costly and timeconsuming to install.

Delineators

Raised pavement markers (RPMs) can be used to show roadway alignment or to replace orsupplement other pavement markings. The same principles that govern the use of paintedmarkings are used for RPMs in terms of color, application, and configuration. The MUTCD (2)provides information on the pattern and spacing of RPMs. The Roadway Delineation PracticesHandbook (16) presents figures to illustrate the principles that the MUTCD outlines and alsospecifically addresses the placement and spacing of RPMs for special situations.

Post-mounted delineators (PMDs) are light-reflecting devices mounted at the side of theroadway, in series, to show the roadway alignment. Their purpose is to outline the edges of theroadway and to accent critical locations. PMDs are usually mounted on posts 1.2 m (4 ft) abovethe pavement. Under normal atmospheric conditions (i.e., no fog, blowing snow, etc.), theyshould be visible at 1000 ft (305 m) when illuminated by the high beams of standard automobileheadlights. In general, PMDs perform best on curves that are 7 degrees or less — for sharpercurves, another form of extra delineation (such as chevrons) should be used. The MUTCD (2)provides standards for the following characteristics:

• mounting height,• number,• spacing,• color of retroreflectors,• criteria for retroreflective elements, and• locations where use is required.

In tangent sections, PMDs should be placed 200 to 500 ft (61 to 153 m) apart in a continuous linenot less than 2 ft (0.6 m) or more than 8 ft (2.4 m) outside the edge of the usable shoulders. Delineators should also be placed on the outside of curves having a radius of 1000 ft (305 m) or


D-11

less. Recommended spacings for delineators on curves are given in the MUTCD. Three PMDsshould be provided both before and after each curve, and the spacing should be such that at leastthree PMDs are always visible to the driver at one time. Generally, the spacing on curves shouldnot exceed 300 ft (91.4 m ) or be less than 20 ft (6.1 m). The Roadway Delineation PracticesHandbook (16) provides information on typical installation of PMD horizontal curves.

Several researchers (26, 36, 37, 38) have reported that post-mounted roadside delineationreduced the crash rate only on relatively sharp curves during periods of darkness. Studies by theArizona Highway Department (39) suggest that neither edgelines nor post-mounted delineationhave any significant effect on the crash rate on open tangent sections.

Other studies indicate that post delineators do have an effect and that highways with postdelineators (in the presence or absence of edgelines) do have lower crash rates than those withoutpost delineators. Further, post delineators are cost justified for all values of cost and service lifefor highways with AADTs exceeding 1000 vehicles per day (40).

Three post-mounted delineator systems used in Virginia were tested in the 1980s at five sites fortheir effectiveness in controlling run-off-road crashes (41). The changes in speed and lateralplacement with the systems in place were taken as driver responses to the systems. The studyindicated that drivers react most favorably to chevron signs on sharp curves greater than or equalto 7 degrees and to standard delineators on curves less than 7 degrees.


D-12

REFERENCES


2. FHWA. Manual on Uniform Traffic Control Devices, Federal Highway Administration,Washington, D.C. (2000).

3. Zegeer, C. V., J. Hummer, D. Reinfurt, L. Herf, and W. Hunter. Safety Effects of Cross-Section Design for Two-Lane Roads, Volumes I and II, Report No. FHWA/RD-87/008,Federal Highway Administration, Washington, D.C. (1987).

4. Glennon, J., T. Newman, and J. Leisch. Safety and Operational Considerations for Designof Rural Highway Curves, Report No. FHWA/RD-86/035, Federal HighwayAdministration, Washington, D.C. (December 1985).

5. Zegeer, C., J. Stewart, F. Council, and D. Reinfurt. Cost-Effective Geometric Improvementsfor Safety Upgrading of Horizontal Curves, Report No. FHWA-RD-90-021, FederalHighway Administration, Washington, D.C. (October 1991).

6. Glennon, J. “Effect of Alignment on Highway Safety,” Relationship Between Safety andKey Highway Features, State of the Art Report 6, Transportation Research Board,Washington, D.C. (1987), pp. 48-63.

7. Fitzpatrick, K., L. Elefteriadou, D. W. Harwood, J. Collins, M. J. McFadden, I. Anderson,R. A. Krammes, N. Irizarry, K. D. Parma, K. M. Bauer, and K. Passetti. Speed Predictionfor Two-Lane Rural Highways. Report FHWA-RD-98-171 (October 1998).

8. Fink, K. L. and R. A. Krammes. “Tangent Length and Sight Distance Effects on AccidentRates at Horizontal Curves on Rural Two-Lane Highways.” Transportation ResearchRecord 1500, Transportation Research Board, Washington, D.C. (1995).

9. Zador, P., H., Stein, J. Hall, and P. Wright. Superelevation and Roadway Geometry:Deficiency at Crash Sites and on Grades (Abridgement), Insurance Institute for HighwaySafety, Washington, D.C. (1985).

10. Harwood, D. W., F. M. Council, E. Hauer, W. E. Hughes, and A. Vogt. Prediction of theExpected Safety Performance of Rural Two-Lane Highways, Report No. FHWA-RD-99-207, Federal Highway Administration, McLean, VA (September 2000).

11. Brinkman, C., and K. Perchonok. “Hazardous Effects of Highway Features and RoadsideObjects Highlights,” Public Roads, Federal Highway Administration, Washington, D.C.(1979), pp. 8-14.


D-13

12. Miaou, S. P. “Vertical Grade Analysis Summary,” unpublished Technical Memorandum toFHWA, (May 1998).

13. Fambro D. B., K. Fitzpatrick, and R. Koppa. “ Determination of Stopping Sight Distances.”NCHRP Report 400, Transportation Research Board, Washington, D.C. (1997).

14. Olson, P. L., D. E. Cleveland, P. S. Fancher, L. P. Kostyniuk, and L. W. Schneider. “Parameters Affecting Stopping Sight Distance.” NCHRP Report 270. TransportationResearch Board, Washington, D.C. (June 1984).

15. Fambro D. B., T. Urbanik II, W. M. Hinshaw, J. W. Hanks Jr., and M. S. Ross. StoppingSight Distance Considerations at Crest Vertical Curves on Rural Two-Lane Highways inTexas. TTI/TxDOT Report 1125-1F. Texas Transportation Institute (March 1989).

16. Migletz, J., J. K. Fish, and J. L. Graham. Roadway Delineation Practices Handbook. Report No. FHWA-SA-93-001. FHWA, U.S. DOT, Washington, D.C. (August 1994).

17. Sign Crew Field Book: A Guide to Proper Location and Installation of Signs and OtherDevices. Texas Department of Transportation. 2nd Edition. (April 1998).

18. Zwahlen, H. T. “Advisory Speed Signs and Curve Signs and Their Effect on Driver EyeScanning and Driving Performance,” Transportation Research Record 1187.Transportation Research Board, Washington, D.C. (1987).

19. Lyles, R.W. An Evaluation of Warning and Regulatory Signs for Curves on Rural Roads.Report RO-80-009. FHWA, U.S. DOT, Washington D.C. (1980).

20. Agent, K. R. Transverse Pavement Markings for Speed Control and Accident Reduction. Kentucky DOT, Lexington, KY (1975).

21. Musiek, J. V. “Effect of Pavement Edge Markings in Two-Lane Rural State Highways inOhio.” Highway Research Board Bulletin 266, Highway Research Board, Washington,D.C. (1962), pp. 1-7.

22. Basile, A. J. “Effect of Pavement Edge Markings on Traffic Accidents in Kansas.”Highway Research Board Bulletin 388, Highway Research Board, Washington, D.C.(1962), pp. 80-86.

23. Bali, S., R. Potts, J. A. Fee, J. I. Taylor, and J. Glennon. Cost Effectiveness and Safety ofAlternative Roadway Delineation Treatments for Rural Two-Lane Highways. ReportFHWA-RD-78-50. National Technical Information Service, Springfield, VA (April 1978).

24. Miller, T. R. “Benefit-Cost Analysis of Lane Markings.” Transportation Research Record1334, Transportation Research Board, Washington, D.C. (1992), pp. 38-45.


D-14

25. Cottrell, B. H. “Evaluation of Wide Edgelines on Two-Lane Rural Roads,” TransportationResearch Record 1160, Transportation Research Board, Washington, D.C. (1988), pp. 35-44.

26. Hall, J. W. “Evaluation of Wide Edgelines,” Transportation Research Record 1114,Transportation Research Board, Washington, D.C. (1987), pp. 21-30.

27. Lum, H. S., and W. E. Hughes. “Edgeline Widths and Traffic Accidents.” Public Roads.Vol. 54, No. 1 (1990), pp. 153-159.

28. Rockwell, T. H., J. Malecki, and D. Shinat. Improving Driver Performance on RuralCurves through Perceptual Changes — Phase III. Ohio DOT, Columbus, OH (1975).

29. Retting, R. A., and C. M. Farmer. “Use of Pavement Markings to Reduce Excessive TrafficSpeeds on Hazardous Curves.” ITE Journal (September 1998).

30. Griffin, L. I., and R. N. Reinhardt. A Review of Two Innovative Pavement Patterns ThatHave Been Developed to Reduce Traffic Speeds and Crashes. AAA Foundation for TrafficSafety, Washington, D.C. (1996).

31. Ullman, G. L. Retroreflective Raised Pavement Marker Field Testing: Initial InterimReport. Report Number 1946-1. (1992).

32. Griffin, L. I. Using the Before-and-After Design with Yoked Comparisons to Estimate theEffectiveness of Accident Countermeasures Implemented at Multiple Treatment Locations. Unpublished Report. Texas Transportation Institute (1990).

33. Kugle, C. L., O. L. Pendleton, and M. S. Von Tess. An Evaluation of the AccidentReduction Effectiveness of Raised Pavement Markers. Report Number TARE 60. TexasTransportation Institute (1984).

34. Mak, K. K., T. Chira-Chavala, and L. I. Griffin. Evaluation on the Safety Effects of RaisedPavement Markers. Unpublished Report. Texas Transportation Institute (1986).

35. Hammond, J. L., and F. J. Wegmann. “Daytime Effects of Raised Pavement Markers onHorizontal Curves.” ITE Journal, Vol. 71, Number 8. (August 2001).

36. Longenecker, K. E. Evaluation of Minor Improvements: Part 1 Delineation. IdahoDepartment of Highways.

37. Tamburri, T. N., et al., Evaluation of Minor Improvements Parts 3 and 4, Delineation andGuardrail. A report of the California Transportation Agency, Sacramento (July 1967).

38. Taylor, J. I., H. W. McGee, E. L. Seguin, and R. S. Hostetter. Roadway DelineationSystems. NCHRP Report 130, Transportation Research Board, Washington, D.C. (1972).


D-15

39. Arizona Highway Department. Delineation vs. Edge Stripe: Cost and Effect. A Report ofthe Arizona Highway Department, Phoenix (June 1963).

40. Capelle, D. G. An Overview of Roadway Delineation Research. Final Report. FHWA-RD-78-111 (June 1978).

41. Jennings, B. E., and M. J. Demetsky. “Evaluation of Curve Delineation Signs,”Transportation Research Record 1010, Transportation Research Board, Washington, D.C.(1985).

APPENDIX E: ROAD SURFACE CONDITION

E-1

Treatments used to decrease crashes associated with wet pavement include filling in pavementruts, improving the skid resistance of a pavement, and warning the motorists of the slipperypavement.

PAVEMENT RUTTING

A study in Wisconsin was conducted to quantify how pavement rutting affects crash rates and toevaluate possible safety-based guidelines for the treatment of pavement rutting (1). Crashes werecategorized as rut-related if the prevailing conditions could be potentially associated with theoccurrence of hydroplanning. Rut depth measurements were average values for both directionsof 1.1-mi (1.8 km) segments and represent the average elevation difference between the tire pathsand the high point between them. The results of the statistical analyses indicated that the definedrut-related crash rate begins to increase at a significantly greater rate as rut depths exceed 0.3inches (7.6 mm). A safety cost-effectiveness curve also demonstrated diminishing marginalreturns when ruts less than 0.3 inches (7.6 mm) are filled. The conclusion was that it iseconomically justifiable in Wisconsin to treat pavements having rut depth measurements of 0.3inches (7.6 mm) or greater.

SKID RESISTANCE IMPROVEMENTS

An example of a site where the skid resistance of a pavement was improved was presented inNCHRP Report 440 (2). A two-lane section of a rural highway located within a state park innorthern California also separates two sections of four-lane freeway. The pavement width variesfrom 24 to 32 ft (7.3 to 9.8 m). The roadway is not a candidate for widening because of sensitiveenvironmental considerations. It is a narrow windy road through an old growth redwood forest(see Figure E-1). The redwoods form a canopy over the roadway which causes the roadway tostay wet and slippery for a while following rain or condensed fog. In addition, the needlesdropping from the trees also contributes to the slipperiness of the roadway. The goal of thetreatment was to reduce wet pavement crashes.

Figure E-1. Two-Lane Rural Highway in an Old Growth Redwood Forest (2).

APPENDIX E: ROADWAY SURFACE CONDITION

E-2

Open graded asphalt concrete (OGAC) has been used by Caltrans for improving wet weather skidresistance and minimizing hydroplaning. Caltrans Standard Specifications currently includesonly a 3/8-inch (0.95 cm) maximum gradation specification. A 1-inch (2.54 cm) maximumgradation provided more voids for better drainage and, thus, better skid resistance by providingmore voids than the 3/8-inch (0.95-cm) or ½-inch (1.27 cm) maximum OGAC standard mix. The 1-inch (2.54 cm) maximum OGAC mix was obtained from the Oregon Department ofTransportation. According to Caltrans, the mix has been used extensively in Oregon and hasbeen successful in reducing the number of crashes. Also, it was used on I-5 where the ADTexceeded 20,000 vehicles, and the pavement has held up well.

A 1-inch (2.54 cm) open graded asphalt concrete was used to reduce wet pavement crashes. Theexisting surfacing was repaired, and dense graded asphalt concrete was placed to level thesurface, especially in two existing pull-out areas. A tack coat was applied to the existing surfaceprior to the placement of the open graded material. The project proposed using a 0.15-ft (0.05 m)thick blanket of the 1-inch (2.54 cm) maximum OGAC on both lanes. The primary purpose of proposing this mix is that the larger amount of voids removes more water, increases traction, andthus reduces the number of crashes.

The estimated cost of the project was $200,000. The work was completed in September 1996. Caltrans believes that the treatment has been performing well. According to their before-and-after study, in the 13 months prior to installation they had 16 wet-pavement-related crashes. They have only had two crashes in the six months after installation. Additional data weregathered as part of this study. Crash data for 32 months prior to installation and 27 monthsfollowing installation were obtained. The average number of crashes before installation was 2.38crashes per month. Following installation, the number dropped to 0.85 crashes per month. Alsonoticeable was the decrease in the number of wet-pavement crashes. Before installation, anaverage of 1.41 wet-pavement crashes per month occurred; after installation, only 0.22 wet-pavement crashes per month occurred. Wet-pavement crashes represented almost 60 percent ofall the crashes on the 2-mi (3.2 km) segment before treatment. After the treatment, they onlyrepresented about 26 percent of the crashes on the segment.

WARNING DEVICES

The MUTCD (3) states that the Slippery When Wet sign may be used to warn that a slipperycondition may exist. When used, a Slippery When Wet sign should be placed in advance of thebeginning of the affected section, and additional signs should be placed at appropriate intervalsalong the road where the condition exists.

FORETELL is being developed by the Federal Highway Administration as part of its ruralintelligent transportation system (ITS) (4). Participants in the program include the statedepartments of transportation for Iowa, Missouri, and Wisconsin. It will provide via the Internettimely, detailed, and relevant weather-related road information needed by state highwaymanagers and the public. The system works by collecting and combining raw weatherinformation from many sources to provide the most recent and accurate weather data available. As of March 2001, FORETELL is in the demonstration phase and can only be accessed by

APPENDIX E: ROAD SURFACE CONDITION

E-3

program partners. Eventually the general public will be able to use FORETELL to access a widerange of weather and pavement condition information for any road or region.

APPENDIX E: ROADWAY SURFACE CONDITION

E-4

REFERENCES

1. Start, M. R., J. Kim, and W. D. Berg. “Potential Safety Cost-Effectiveness of TreatingRutted Pavements.” Transportation Research Record 1629. Paper 98-0465 (1998).


3. FHWA. Manual on Uniform Traffic Control Devices. Federal Highway Administration,Washington, D.C. (2001).

4. Pisano, P. “FORETELL – Finally, Someone Is Doing Something about the Weather.”Public Roads (March/April 2001).

APPENDIX F: NARROW BRIDGE

F-1

Highway bridges are sometimes associated with crash problems, particularly rural highwaybridges with narrow width, poor sight distance (e.g., just past a sharp horizontal curve), and/or poor signing and delineation. Numerous studies have analyzed the effects of various trafficcontrol devices (e.g., signs and markings) on crashes and on vehicle operations such as vehicleplacement on the bridge. However, research is scarce on the effects of bridge geometrics oncrash experience.

The features, which are of the most importance in terms of affecting bridge crash rates are thebridge width and/or the width of the bridge in relation to the approach width. The best knowncrash relationship with bridge width was developed in a 1984 study by Turner (1). Based oncrashes at 2087 bridges on two-lane roads in Texas, a crash model was developed as a functionof “relative bridge width” (RW), which is defined as the bridge width (C) minus the width of thetraveled way (B) (see Figure F-1).

According to Turner’s crash model, as shown in Figure F-2, the number of crashes per millionvehicles decreases as the relative bridge width increases (1). This relationship indicates that it isdesirable to have bridge widths at least 6 ft (1.8 m) wider than the traveled way. In other words,shoulders of 3 ft (0.9 m) or more should be provided on each side of the bridge.The relationship shown in Figure F-2 is currently the best information available on the topic;however, the reader should note that the study did not include bridges with no crashes. If thesebridges would have been included, a different relationship may have been found.

Where: A = Lane WidthB = Traveled Way WidthC = Bridge WidthD = Approach Roadway WidthRW = Relative Bridge With RW = Bridge Width (C) - Traveled Way Width (B)

Figure F-1. Key Elements at a Bridge Site (1).


F-2

0.0

0.2

0.4

0.6

0.8

1.0

-10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16

Relative Bridge Width, RW(ft)

Acc

iden

ts p

er M

illio

n V

ehic

les

Y = 0.50 - .061(RW)0.0022(RW)2

RW = Relative Bridge Width

= Bridge Width - Traveled Way Width

1 ft = 0.305 m

Figure F-2. Crash Rate by Relative Width (1).

Listed below are potential countermeasures identified as a means of reducing crash frequencyand severity at bridges.

UPDATE BRIDGE RAILING

All bridge structures are required to have some type of bridge railing in place to prevent vehiclesfrom running off the edge of a bridge or culvert. Bridge railing differs from roadside barriers inthat they are generally an integral part of the structure (i.e., physically attached), and they aredesigned to have virtually no deflection when struck by an errant motorist.


F-3

According to the Roadway Design Guide (2), bridge rails designed to AASHTO specificationsprior to 1964 may not meet current specifications and may need retrofitting. Retrofit designsmay be needed to do the following:

� Increase the strength of the railing system.� Provide longitudinal continuity.� Reduce or eliminate undesirable effects of curbs or narrow walkways in front of the bridge

rail.� Eliminate snagging potential.� Increase the height of the rail systems to accommodate higher profile vehicles.

Several options exist for retrofitting or updating bridge rail systems, and the reader shouldconsult the Roadway Design Guide for more information on the design and implementation ofthese options.

INCREASE BRIDGE WIDTH

Although expensive, increasing the width of a bridge is another option available for correctingcrash problems associated with bridges. It is desirable that a bridge be designed to provide a full,continuous shoulder so that a uniform clearance to the bridge sides is maintained. The uniformalignment created by maintaining the full shoulder widths enhances highway safety by reducingdriver concern for and reaction with the structural elements of the bridge rail.

Crash reduction factors given in Table F-1 provide percent reductions in total crash rate expecteddue to widening shoulders on bridges. For example, assume that a bridge is 24 ft (7.3 m) widewith 10-ft (3.1 m) lanes and 2-ft (6 m) shoulders on each side. According to Table F-1, wideningthe bridge to 32-ft (9.8 m) (i.e., two 10-ft [3.1 m] lanes with two 6-ft [1.8 m] shoulders) wouldreduce the total bridge crash rate by 62 percent.

Note that values in Table F-1 assume that the lane width stays constant in the before-and-aftercondition. When the bridge lane width is increased, a conservative estimate of crash reductionwould be to use Table F-1 and only include the amount of increased shoulder width. Forexample, when widening a 20-ft (6.1 m) bridge (two 10-ft [3.1 m] lanes and no shoulder) to a 30-ft (9.1 m) bridge (two 12-ft (3.7 m) lanes and two 3-ft (.9 m) shoulders), assume an increasein shoulder width from 0 to 3 ft (0 to .9 m) for at least a 42 percent “minimum” crash reduction.While the factors shown in Table F-1 are the best estimate available of the effect of bridge widthon crashes, they should be used cautiously because of an important drawback of the study onwhich they are based. This study considered only bridges that experienced one or more crashesduring the study period. Failure to include sites that experienced no crashes is a well-knownsource of bias in safety research.


F-4

Table F-1. Summary of Crash Reduction Factors Associated with WideningShoulders on Bridges (1). a

Bridge ShoulderWidth Before

Widening, m (ft)

Bridge Shoulder Width, m (ft) after Widening Each Side[Total of Both Sides in Brackets]

EachSide

Total ofBothSides

1.2 [0.6](4 [2])

1.8 [0.9](6 [3])

3.1 [1.2](10 [4])

2.4 [1.5](8 [5])

3.7 [1.8](12 [6])

4.3 [2.1](14 [7])

4.9 [2.4](16 [8])

0 (0) 0 (0) 23 42 57 69 78 83 85

1 (0.3) 2 (0.6) -- 25 45 60 72 78 80

2 (0.6) 4 (1.2) -- -- 27 47 62 71 74

3 (0.9) 6 (1.8) -- -- -- 28 48 60 64

4 (1.2) 8 (2.4) -- -- -- -- 44 44 50a Assume that the width of lanes on the bridge remains constant. Values in the table were derived based on thecrash model developed by Turner on rural two-lane roads.

IMPROVE SIGNING AND DELINEATION

Controlled field studies were used to examine whether the amount and type of delineationprovided at different bridge/culvert designs had an impact on drivers’ comfort levels whenapproaching and crossing narrow bridges (3). Although researchers did not find a delineation technique that performed significantly better than the rest studied, they included recommendedtapered edgeline/transverse marking arrangements. Figure F-3 shows the recommendation forroadways with edgelines, an offset-bridge clearance < 3.3 ft (1 m), and no lane width reductionacross the bridge. They also included drawings for the following two cases: 1) roadways withedgeline offset > 3.9 ft (1.2 m) and offset bridge clearance > 3.3 ft (1 m) and 2) roadways withoutedgelines and approach width greater than the bridge width. The configurations arerecommended for narrow bridge locations where a crash problem is known to exist or whereother evidence suggests that drivers are not vacating the shoulder soon enough to avoid strikingthe bridge.

EXAMPLE OF A TREATMENT ON A NARROW BRIDGE

NCHRP Report 440 (4) discussed a treatment used on a narrow bridge in Missouri. A series ofcrashes prompted local residents to become concerned about the safety of a bridge located on atwo-lane rural highway in the midwest (see Figure F-4). The bridge, carrying an average dailytraffic of approximately 8000 vehicles with 7 percent trucks, was located on a highway that wasbecoming a premier route for truck drivers. Primarily, safety problems involved the narrowwidth of the bridge and the growing truck traffic. Specifically, the following situations wereencountered at the bridge:

• drivers stopping at the bridge when a large vehicle was crossing, causing rear-end crashes;• drivers encroaching the centerline of the bridge, causing sideswipe crashes; and


F-5

• left-turn traffic at lake roads located at the end of the bridge approaches, contributing to rear-end crashes.

Notes:• Type 2 or Type 3 object marker.• Type object marker if guardrail protection for bridge end is provided.• Optional PMDs at 24.93- to 49.87-ft (7.6 to 15.2 m) spacings.

Figure F-3. Recommended Bridge Delineation for Roadways with Edgelines, Offset-BridgeClearance of < 3.3 ft (1 m), and No Lane Width Reduction Across Bridge (3).

Figure F-4. Narrow Bridge (2).


F-6

Concerned citizens, organized by one of the motorists injured at the bridge, urged the departmentof transportation to replace the bridge. Lacking sufficient funds until a much later date, thedepartment of transportation researched new products on the market and the latest technologyapplicable to this situation. Combining several different products to fit the situation, they createdthe Electronic Advanced Warning System currently in place. The Electronic Advanced WarningSystem was designed to 1) warn motorists of large vehicles crossing the bridge in the opposinglane and 2) warn motorists upstream of the bridge of vehicles stopped at the bridge.

Motorists are warned of large vehicles approaching in the opposing lane by a flasher on top of a“Caution” sign located at both ends of the bridge. Narrow beam microwave units, mounted on apost at either end of the bridge, detect vehicle direction and height. These units are set to detectapproaching vehicles with a minimum height of 10 ft (3.1 m). Radio telemetry is utilized toactivate the flasher on the opposite side of the bridge.

The second part of the Electronic Advanced Warning System was the installation of a “BePrepared to Stop” sign with a flasher about 1200 ft (366 m) from both ends of the bridge. Usinginduction loop technology, an 80-ft (24.4 m) detection zone was created to detect stopped orslow-moving traffic. When these criteria are met, the flasher on the sign begins to flash, warningapproaching traffic of the condition ahead.

The work was completed in February 1996 at an approximate cost of $33,000. This equipment,although somewhat susceptible to failure from lightning, has worked very effectively. Themotorists seem to understand the function of the equipment and heed its warning. The trafficsection believes this is an effective short-term solution with logical expectations and benefits. The department of transportation estimated the annual crash reduction realized by thiscountermeasure. These estimates are summarized in Table F-2.

Table F-2. Annual Crash Reduction for Treatment at Narrow Bridge (4).

Crash Type

EstimatedReduction

Annual Number CrashesBefore Improvement

Estimated Annual Reductionin Crashes

% PDO Fatal/Injury PDO Fatal/Injury

Side SwipeRear EndAvoiding

Out of Control

24242424

4.20.81.20.4

1.21.00.20.8

1.000.190.290.10

0.290.240.050.19

Total Estimated Crash Reduction: 1.58 0.77


F-7

REFERENCES

1. Turner, D. S. “Prediction of Bridge Accident Rates,” Journal of Transportation Engineering,Vol. 110, No. 1, American Society of Civil Engineers, New York, NY (January 1984), pp. 45-54.

2. AASHTO, Roadside Design Guide, American Association of State Highway andTransportation Officials, Washington, D.C. (1996).

3. Ullman, G. L., and V. J. Pezoldt. Delineation of Bridges and Culverts in Texas. FHWA/TX-98/1366-1F. (February 1997).

4. Fitzpatrick, K., D. W. Harwood, I. B. Anderson, and K. Balke. Accident Mitigation Guide forCongested Rural Two-Lane Highways. NCHRP Report 440. Transportation Research Board,Washington, D.C. (2000).

APPENDIX G: INTERSECTIONS

G-1

A major safety concern in rural areas is the result of speed differentials, generally due to vehiclesslowing or stopped to make a turning maneuver at the intersection. For minor rural intersections,the effects of speed differential may be reduced greatly by flaring the intersection and permittingthrough traffic to bypass to the right of the vehicle waiting to make a left turn. Major ruralintersections may need turn bays as a means of maintaining high operating efficiency and a saferenvironment. When a large number of turns need to be accommodated or when the intersectionarea is large and vehicles need additional guidance through the intersection, channelization maybe considered.

Conflicts between turning vehicles and pedestrians and between turning vehicles and othervehicles approaching from the opposite direction can cause congestion delay and safety problemsat intersections and driveway access points. Hummer et al. (1) used crash data for 1993 to 1995for two selected counties in North Carolina to determine the types of collisions typicallyassociated with rural intersections and driveways (see Table G-1). They found that rear-endcrashes, which may include crashes involving turning vehicles, were the most common. Crashesinvolving left- or right-turn maneuvers represented 33 percent of the passenger car crashes and34 percent of the truck crashes.

Table G-1. Collision Types at Rural Intersections (1).

Collision Type Car Truck

Number Percent Number Percent

Rear EndAngle

Left TurnRight Turn

BackingOther

13111249115317711316

333129430

371345348605110

312929541

TOTAL 4019 100 1185 100

An Ohio Department of Transportation study (2) determined the crash rates at unsignalized andsignalized intersections with and without turn lanes. Table G-2 shows the significant differencesin the crash rates for the different categories.

A 1975 Kentucky study (3) at yield-controlled intersections indicated that over half the crasheswere rear-end collision while angle collisions were over half the crashes at stop signs (see TableG-3). Table G-3 also lists the results from a similar 1976 study, which used data from ruraltowns in Virginia but did not differentiate between crashes at yield signs and those at stop signs. The Virginia study also noted that crash rates at stop-controlled intersections were lower at thoseintersections having high traffic volumes (see Table G-4).


G-2

Table G-2. Crash Rates for Intersections with Signal Control and Left-Turn Lanes (2).

Type of CrashUnsignalized Signalized

No Left-TurnLane

With Left-TurnLane

No Left-TurnLane

With Left-TurnLane

Left TurnsAll Other

1.20 3.15 S

0.12 0.92 S

0.65 1.82 S

0.37 1.17

TOTAL 4.35 S 1.04 S 2.47 S 1.54 S

Crashes per million entering vehicles“S” denotes a difference that is statistically significant.

Table G-3. Crash Types at Sign-Controlled Rural Intersections (3).

State Control Percent of All Crashes Crash Rate (crashes per millionentering vehicles)

Rear End orSideswipe

RightAngle

Other

Kentucky Yield SignsStop Signs

56.229.6

22.551.9

21.318.5

Not available

Virginia Stop andYield signs 39 49 12 1.08

Table G-4. Relationship of Crash Rates to Traffic Volume Entering Stop-Controlled Intersections (4).

ADT Crash Rate (crashes per million entering vehicles)

Less than 10,00010,000 to 15,00015,000 to 20,000

Over 20,000

1.121.050.970.52

LEFT-TURN LANE

The left-turn lane is generally the key auxiliary lane at an intersection. It creates the opportunityto separate and avoid speed differences between the turning vehicle and the through vehicles. Italso decreases the delay that can be experienced by through vehicles behind a turning vehicle. By increasing the operational efficiency of the intersection, the capacity and safety are alsoincreased. In addition, left-turn lanes can provide increased visibility to the turning vehicle bythe opposing traffic. It can also increase the distance downstream that the turning driver can seeand will result in the driver being able to better judge the availability of gaps.


G-3

The AASHTO Green Book (5) indicates that left-turn lanes should be established on roadwayswhere traffic volumes are high enough or safety considerations are sufficient to justify left-turntreatment. Green Book Exhibit 9-75 lists the traffic volumes where left-turn lanes should beconsidered. Additional information on left-turn treatments at intersections is included in NCHRPSynthesis 225 (6) and NCHRP Report 279 (7).

Providing adequate deceleration into a turn bay at an intersection will result in the smoothremoval of a turning vehicle from the through lane. When the deceleration distance is too short,vehicles desiring to turn will be at a lower speed in the through lane (as they decelerate inanticipation of the turn bay) than the through vehicles. This differential in speeds can causeconflicts and crashes. Speed differential can also occur between a vehicle accelerating after a turnand the through vehicles in the same lane. Adequate deceleration and acceleration transitionareas can improve operations and safety. Information on taper designs anddeceleration/acceleration lengths for different grades or running speed assumptions is included inthe Green Book (5).

The storage requirements for turn lanes is computed on the basis of the number of vehicles to bestored. The storage length should be sufficient to avoid the possibility of left-turning vehiclesstopping in the through lanes. The storage length should also be sufficiently long so that theentrance to the auxiliary lane is not blocked by vehicles standing in the through lanes waiting fora signal change or for a gap in the opposing traffic flow.

At unsignalized intersections, the storage length, exclusive of taper, may be based on the numberof turning vehicles likely to arrive in an average two-minute period within the peak hour (5). Asa minimum requirement, space for at least two passenger cars should be provided; with over 10percent truck traffic, provisions should be made for at least one car and one truck. The two-minute waiting time may need to be changed to some other interval that depends largely on theopportunities for completing the left-turn maneuver. These intervals, in turn, depend on thevolume of opposing traffic.

A California study examined the difference in the effectiveness of the raised barrier protected leftturn versus the painted left turn in rural areas (8). Both treatments provided a significantreduction in crash rates with relatively little difference between the types of treatment for ruralareas (see Table G-5). The study’s findings for the urban intersections indicated that the raisedbarrier protected left-turn lanes were much more effective than painted left-turn lanes. The studyalso compared crash reduction resulting from adding left-turn channelization at signalized andunsignalized intersections. The signalized intersections experienced an 18 percent reduction(from 1.00 to 0.82 crashes per million entering vehicles) while the unsignalized experienced a 50percent reduction (from 1.16 to 0.58 crashes per million entering vehicles).


G-4

Table G-5. Crash Rates Before and After Adding Left-Turn Channelization at Unsignalized Intersections in Rural Areas (8).

Raised Barrier Protected Left-Turn Lane

Painted Left-Turn Lane

RateBefore

Rate After PercentChange

RateBefore

Rate After PercentChange

Crash TypeSingle VehicleLeft TurnRear EndCrossingOther

0.100.180.490.280.13

0.070.050.020.270.07

-30 -72 -96 S- 4 -46

0.100.280.510.190.07

0.150.150.090.160.03

+50 -46 -82 S-16 -57

SeverityProperty DamageInjuryFatal

0.720.390.08

0.340.150.00

-53 S-62 S-100

0.610.540.01

0.310.250.01

-49 S-54 S

0

Light ConditionDayNight

0.670.51

0.250.24

-64 S-53 S

1.181.13

0.550.63

-53 S-44

TOTAL 1.18 1.049 -58 S 1.16 0.58 -50 S

Changes indicated with “S” are significant at the 0.10 level using the chi-square test.Crash rates are the number of crashes per million entering vehicles.

NCHRP Report 440 (9) presented a site where left-turn treatments were added to a two-laneroadway that has several world-class vineyards and restaurants that attract tourists. The roadwayis a two-lane highway with shoulders on the roadway that are generally 4 to 8 ft (1.2 to 2.4 m)wide with a peak-hour volume of 2200 vehicles in 1983. The roadway is particularly congestedon weekends because of the high number of visitors to the wineries. Queues often form behindvehicles waiting to make a left turn into the wineries. During the period between September 1,1983, and August 31, 1986, there were 138 rear-end or left-turn-related crashes.

The construction project widened the roadway to include left-turn lanes at selected intersections(see Figure G-1) and a two-way turn lane on the north end of the project (see Figure G-2). Thepavement was rehabilitated and widened to a 40-ft (12.2 m) cross section that included two 12-ft(3.66 m) lanes and two 8-ft (2.4 m) shoulders. In the vicinity of several intersections, theexisting pavement was widened to 52 ft (15.9 m) to accommodate two 8-ft (2.4 m) shoulders,two 12-ft (3.7 m) through lanes, and one 12-ft (3.7 m) left-turn lane. In addition, all existingleft-turn bays were brought up to the 52-ft (15.9 m ) dimension. The improvements wereanticipated to decrease the number of left-turn-related crashes in addition to reducing congestionand delay.


G-5

The work was completed in June of 1996 at a cost of $3.2 million. Crash data for the roadwaysegment were obtained for 29 months prior to installation and 30 months following installation. Before the left-turn treatments were installed, an average of 2.41 crashes per month occurred. Following installation, an average of 1.83 crashes per month occurred, representingapproximately a 24 percent reduction in number of crashes. The type of primary collision factorassociated with the crashes remained fairly constant between the two periods. The most commonprimary collision factor was speeding (36 percent of the before crashes and the after crashes),followed by failure-to-yield (19 percent of the before crashes and 15 percent of the after crashes),and improper turn (9 percent of the before crashes and 15 percent of the after crashes). It appearsthat the type of crashes did not change between the two periods; however, the total number ofcrashes did decrease.

Figure G-1. Example of Left-Turn Figure G-2. Example of Two-Way TurnLane (9). Lane (9).

SHOULDER BYPASS LANES

Shoulder bypass lanes are a low-cost alternative to intersection turn lanes for reducing delays tothrough vehicles caused by left-turning vehicles. Where a side road intersects a two-lanehighway at a three-leg or T-intersection, a portion of the paved shoulder opposite the intersectionmay be marked as a lane for through traffic to bypass vehicles making a left turn. The bypasslane may also be used at major driveways. Where an adequate paved shoulder is alreadyavailable, installation of a shoulder bypass lane may be as simple as remarking the highwayedgeline. Thus, provision of a shoulder bypass lane is often much less expensive thanconstruction of a left-turn lane. At other locations, construction of a paved shoulder for use as abypass lane may be justified either to improve traffic operations or reduce crash experience.

Figure G-3 illustrates a typical shoulder bypass lane at a T-intersection on a two-lane highway. Ifa vehicle is stopped in the through travel lane waiting to make a left turn, following vehicles canuse the bypass lane to avoid having to stop themselves. The marking of a bypass lane encouragesdrivers to avoid unnecessary delay and assures that the maneuver is legal by designating a portionof the paved shoulder as part of the traveled way.


G-6

1 ft = 0.305 m

Figure G-3. Plan View of Typical Intersection with Shoulder Bypass Lane (10).

Shoulder bypass lanes have been shown to be effective in reducing delay to through vehicles atT-intersections as well as reducing fuel consumption, vehicle operating costs, and pollutantemissions. No quantitative estimates are available for the delay reduction effectiveness ofshoulder bypass lanes. However, a Delaware study found that, where shoulder bypass lanes areprovided, 97 percent of the drivers who needed them to avoid delay did in fact use them (11). Similarly, an Illinois study observed over 90 percent usage of shoulder bypass lanes by driverswho needed them (12). Even bypass lanes as short as 150 ft (46 m) were used effectively bydrivers.

Shoulder bypass lanes were found to be more effective than paved shoulders alone in improvingtraffic operations. In Delaware, where use of both paved shoulders and shoulder bypass lanes tobypass left-turning vehicles is legal, only 81 percent of drivers used paved shoulders to bypassleft-turning vehicles, whereas 97 percent of drivers used shoulder bypass lanes where necessary.

The crash experience of shoulder bypass lanes compared with that of separate left-turn lanes orcompared with that of paved shoulders alone has not been formally evaluated. However,Nebraska has reported a marked decrease in rear-end crashes at shoulder bypass lanes, and otherstates have reported relatively few crashes occurring at shoulder bypass lane installations (11).

RIGHT-TURN LANE

Right-turn lanes can provide increased operational efficiency to an intersection, especially whenthere is a high volume of vehicles turning right. The design of the right-turn lane is very similarto the design of the left-turn lane. There should be adequate storage and a smooth taper into thelane.

Several publications provide advice on when to consider a right-turn lane. Some of theinformation is based on a benefit-cost analysis or on the safety effects of the turn lane. Most use


G-7

operational effects to identify when a turn lane should be considered. Stover et al. (13) suggestthat right-turn lanes be provided on uncontrolled intersection approaches when the average dailytraffic on the intersecting roadway is 500 vehicles per day or greater. Glennon et al. (14)conducted a benefit-cost analysis and made assumptions about the operational and safety effectsof right-turn lanes. The results of the analysis indicated that right-turn lanes are cost-effective atdriveways when a) the driveway volume is at least 1000 vpd with at least 40 right turns into thedriveway during peak periods, and b) the roadway ADT is at least 10,000 vpd and the roadwayspeed is at least 35 mph (56 km/h). Cottrell (15) developed guidelines using informationobtained from a survey of state practices and field studies. The treatments considered were: a) nospecial treatment other than the radius, b) a taper, and c) a full-width lane.

McCoy et al. (16) developed guidelines for the use of right-turn lanes at access points on urbantwo-lane and four-lane roadways. The study was performed for the Nebraska Department ofRoads. The guidelines compared the benefits and costs of right-turn lanes at uncontrolledintersections and driveways on urban roadways. Benefits included cost savings in delay, fuelconsumption, and crashes. Costs considered construction and maintenance of the turn lane.Table G-6 lists the guidelines for urban two-lane roadways.

Hasan and Stokes (17) also developed guidelines for right-turn treatments at unsignalizedintersections and driveways on rural highways. Their guidelines considered two types oftreatments: full-width lane and taper. A benefit-cost evaluation along with operational effectsand safety effects were considered during the development. The safety effects evaluation usedthe relationship between speed differential and crash to estimate the reduction in right-turn,same-direction, rear-end crashes that would be expected to result from the provision of a right-turn treatment. Table G-7 lists the guidelines developed.

Table G-6. Right-Turn Lane Guidelines for Urban Two-Lane Roadways (16).

Road-way

DDHV(vph)

Minimum Right-Turn DHV (vph)

Within ExistingROW

ROW Cost =$0.093/m2

ROW Cost =$0.465/m2

ROW Cost =$0.93/m2

Roadway Speed(km/h)

Roadway Speed(km/h)

Roadway Speed(km/h)

Roadway Speed(km/h)

40 56 72 89 40 56 72 89 40 56 72 89 40 56 72 89

10012515020040060080010001200

-6560504035302525

-6050453530252020

654035302015151515

302520151010101010

-7065554035303030

-6555453530252525

705040302015151515

402520151010101010

--

75654035303030

--

75654035303030

-7560403025202020

-4535252015101010

--

95805545353535

--

95805545353535

--

90604035303030

--

50302015151515

* 1 km/h = 0.6 mph


G-8

Table G-7. Right-Turn Treatment Guidelines for Two-Lane Highways (17).

Road-way

DDHV(vph)

Minimum Right-Turn DHV (vph) a

Lane Taper

Roadway Speed (km/h) Roadway Speed (km/h)

64 72 81 89 97 105 64 72 81 89 97 150

20030040060080010001200

20050251410

120522616129

73413020151198

352419141198

20151210987

1512119877

85271286

83402713854

3019149754

14986544

8765433

7654333

a Minimum right-turn design hour volume (vph) required to warrantright-turn treatments based on an assumed turning speed of 24 km/h.1 km/h = 0.6 mph

CHANNELIZATION

Potential conflicts among vehicles and between vehicles and pedestrians may be reduced throughchannelization of traffic movements. The traffic may be channeled into specific and clearlydefined vehicle paths. Operational objectives of channelization are as follows (18):

• direct traffic movements,• assure orderly movement,• increase capacity,• improve safety,• maximize effective traffic control and communication with the driver, and• reduce conflicts.

Because traffic volumes, pedestrian patterns, and physical conditions vary, individualchannelization treatments are generally needed for each intersection. Good design should adhereto the following principles (18):

• The proper traffic channels should appear natural and convenient to drivers and pedestrians. There should be no choice of vehicle paths leading to the same destination. The number ofislands should be held to a practical minimum to avoid confusion.

• Islands should be large enough to be effective. Islands that are too small are ineffective as amethod of guidance and often present problems in maintenance. The area of an island shouldbe at least 75 ft2 (7 m2). Accordingly, triangular islands should not be less than about 12 ft(3.7 m) on a side, after the rounding of corners. Elongated or divisional islands should be atleast 4 ft (1.2 m) wide and 12 to 20 ft (3.7 to 6.1 m) long.


G-9

• Channelization should be visible. It should not be introduced where sight distance is limited. When an island must be located near a high point in the roadway profile or near thebeginning of a horizontal curve, the approach end of the island should be extended so that itwill be clearly visible to approaching drivers.

• The major traffic flows should be favored.• Conflicts should be separated so that drivers and pedestrians may deal with only one conflict

and make only one decision at a time.• Islands should be designed for the design speed of the road. The approach end treatment and

delineation should be carefully designed to be consistent with the speed characteristics of theroadway design.

Additional guidance on channelization is provided in several reference materials, such as theGreen Book (5), Stover and Koepke (18), and NCHRP Report 279 (7).

The effectiveness of various safety improvement projects was evaluated in the early 1970s byDale (19) . He found that channelization of intersections produced an average 32.4 percentreduction in all types of crashes. Crashes involving personal injuries decreased by over 50percent. An analysis done in 1978 by Strate (20) of the impact of 34 types of safety improvementprojects indicated that intersection channelization projects had produced an average benefit/costratio of 2.31.

RUMBLE STRIPS ON APPROACHES TO INTERSECTIONS

While the TMUTCD (21) does not provide information on the use of rumble strips at anintersection, it does provide the following support for transverse rumble strips as temporarytraffic control devices and as approach end treatments for curb islands.

Rumble strips consist of intermittent narrow, transverse areas of rough-textured orslightly raised or depressed road surface that alert drivers to unusual motor vehicle trafficconditions. Through noise and vibration they attract the driver’s attention to suchfeatures as unexpected changes in alignment and to conditions requiring a stop.

It also provides the following options and guidance:

• Intervals between rumble strips may be reduced as the distance to the approached condition isdiminished in order to convey an impression that a closure speed is too fast and/or that anaction is imminent. A sign warning drivers of the onset of rumble strips may be placed inadvance of any rumble strip installation.

• Rumble strips should be placed transverse to motor vehicle traffic movement. They shouldnot adversely affect overall pavement skid resistance under wet or dry conditions.


G-10

• In urban areas, even though a closer spacing may be warranted, care should be taken not topromote panic braking or erratic steering maneuvers by drivers.

• Rumble strips should not be placed on sharp horizontal or vertical curves.

ADVANCE WARNING

Studies have examined the effect of different signs and beacons on crashes or speed. A regulatoryspeed-zone configuration and lighted warning signs were found to be more effective than moretraditional unlighted warning signs in reducing motorists’ speeds in the vicinity of a ruralintersection and increasing their awareness of both the signs and the conditions at the intersection(22). A study of rural high-speed intersections found that providing the driver with adequatewarning of the intersection is of primary importance for this type of intersection (23).

NCHRP Report 440 (9) included an example of where a flasher was used on an advance warningsign along with other treatments. The intersection of two state highways has experienced severalsevere side and rear-end collisions over the past 25 years. Initially, the intersection was a two-way stop on the minor cross street. Historically, daily travel volumes on the major highway havebeen double the volumes of the minor roadway. The intersection is in a rural area with limiteddevelopment surrounding the intersection.

Several approaches have been implemented at this location to improve safety with limitedresults. In November 1980, overhead flashing signals were installed at the intersection. Aflashing red beacon was installed on the minor highway, and yellow flashing beacons wereplaced overhead for the major approach. Officials noticed improvement for the intersection, butas traffic volumes continued to increase, crashes directly associated with the intersectionincreased. One potential concern with the overhead beacons was that unfamiliar drivers did notreceive adequate decision time because of the complex and unexpected intersection on arelatively straight rural roadway. In September 1994, right-turn lanes were added to all fourapproaches with “Cross Street Does Not Stop” advanced warning signs on the minor approach. Recent severe crashes prompted a review of the location. In December 1997, a decision wasmade to remove the overhead flashers and to convert the intersection to all-way stop control.

In addition to converting the control at the intersection to an all-way stop, additional signing andbeacons were added. Advance red flashing beacons with stop ahead signs were installed on allapproaches (see Figure G-4). At the intersection, flashers were added to the stop signs (seeFigure G-5). The project was completed in March of 1998. State officials state that the next steptoward improving this location would be to install signal heads on all approaches. Currently,traffic volumes do not meet MUTCD warrants for a traffic signal.

Preliminary results suggest that the countermeasure has been effective at reducing the number ofcrashes. Interviews conducted with state traffic operations officials and local store ownersindicated that there have not been any crashes at that location since installation of the advancewarning signs and the flashing beacons, and that the treatment is well received.


G-11

Figure G-4. Beacons on Stop Ahead Sign (9). Figure G-5. Beacons on

Stop Sign (9).

SUPPLEMENTAL SIGNS ON STOP SIGNS

The traffic control for low-volume rural intersections is generally no-control, yield, or stop signs. A 1978 NCHRP study (24) made the following comments on the general safety aspects of signcontrols at intersections:

• Yield signs effectively reduce crashes at low-volume, isolated urban intersections.• Four-way stop controls significantly reduce crashes at intersections where entering traffic

volumes on all approaches are relatively equal.• Four-way stop controls result in increased crashes where traffic volumes on approaches are

not relatively equal.

Signs warning motorists that traffic on the cross street does not stop can be found at someintersections that are not all-way stop controlled. These “cross traffic” signs have been installedto provide a special warning where some motorists on the minor approach may incorrectlyassume that the major crossing street also has stop signs. A review of crash data offered mixedresults about the signs’ effectiveness: at some locations, the signs appeared to reduce crashfrequencies; at others, crashes continued despite their presence (25). Care should be taken tocontrol the use of the sign because expanded use could cause drivers to expect them at all two-way, stop-controlled situations. More information on the long-term impact of the signs is needed. Another study determined that if supplemental signs are used, most drivers understand and preferthe design shown in Figure G-6.


G-12

Figure G-6. Preferred Supplemental Sign at Two-Way, Stop-Controlled Intersections (26).

INTERSECTION FLASHING BEACON

Intersection control beacons have flashing yellow or red indications on each face. They areinstalled and used only at an intersection to control two or more directions of travel. They areintended for use at intersections where traffic or physical conditions do not justify conventionaltraffic signals but where high crash rates indicate a special hazard. Intersection control beaconsare used in conjunction with stop signs at isolated intersections or intersections having sightdistance obstructions.

Results of a 1970 North Carolina State University study (27) of crashes before and afterinstallation of flashers at stop sign controlled rural intersections are shown in Table G-8. Theauthors state that there was a statistically significant decrease in crash rates on the aggregatesites, on three and four sections. Most noticeable was the decrease in single-vehicle crashes.

Table G-8. Change in Crash Experience with Addition of Flashers at Stop Sign Controlled Rural Intersections (27).

Intersection TypePercent Change

Total SingleVehicle

Left-Turn Rear-End Angle Other

4 Leg3 Leg

ChannelizedNon-channelized

- 18- 65- 4724

- 62- 62 - 63- 50

- 24-

+70+1

- 5- 100- 63

3

- 18- 100- 5088

- 4- 50- 33 32

TOTAL -27 -62 -13 -33 -21 -17

Results of a similar study in California (8) of changes in crash patterns as a result of installationof flashing beacons at stop sign controlled intersections was summarized as:

• Total crashes decreased 43 percent.• Single-vehicle crashes decreased 67 percent.• Left-turn crashes decreased 39 percent.


G-13

• Rear-end crashes decreased 17 percent.• Angle crashes decreased 45 percent.• Other two-vehicle crashes decreased 47 percent.

The severity of crashes was also reduced:

• Property damage crashes decreased 34 percent.• Injury crashes decreased 51 percent.• Fatal crashes decreased 80 percent.

There was a marked decrease in both daytime and nighttime crashes; those in the day decreased43 percent, those at night decreased 46 percent.

Table G-9 shows a comparison of safety impacts for different types of flasher control. It isinteresting that the addition of four-way red flashers has an effect somewhat similar to that oftraffic signal control: that angle collisions are reduced but rear-end crashes increase significantly. The decrease in severity of crashes and in the number occurring in daytime and nighttime hourswas quite similar to the averages previously described for all crashes.

Table G-9. Change in Crash Rates at Intersections with Addition of Flashing Beacons (8).

Crash TypePercent Changes

Red-Yellow Flashers 4-Way Red Flashers

3-Leg 4-Leg

Single Vehicle - 29 - 82 - 52

Multiple Vehicle

Left TurnRear EndAngleOther

- 7- 46- 33- 25

- 44- - - 14- 63

- 82100- 82- 73

- - = No crashes occurred in the before period.

Table G-10 indicates that the California study did not find a significant difference in effectbetween flashers that were installed at channelized intersections and those at non-channelizedintersections. An interesting facet of the California study was a comparison of the impact oncrash rates produced when a four-way red flasher (i.e., four-way stop control) was installed atintersections with various previous forms of traffic control as shown in Table G-11.

The California study also analyzed the before-and-after severity of crashes, as a result ofinstalling flashing yellow beacons at the approaches of intersections. While there was an


G-14

increase in personal injury crashes, property damage crashes decreased 41 percent, and there wasa 100 percent decrease in fatalities.

Table G-10. Change in Crash Rates with Red-Yellow Flashing Beacons Added AtChannelized and Non-Channelized Intersections (8).

Channelization Present Percent Change

3-Leg 4-Leg

ChannelizedNon-Channelized

- 51- 54

- 25- 38

A study in Ohio examined 82 intersections, each of which was controlled by a flashing beacon(28). The results indicated that there is a reduction in crash rate with the installation of a flashingbeacon. The evaluation of the different types of flashers revealed that intersections had asignificant reduction in total crashes when equipped with the following types of flashers: 1)standard stop sign on the side of the road with one or two flashing beacons attached to thesupport post; 2) a single unit placed overhead in the center of the minor approach roadway anddisplaying two beacons flashing alternately; and 3) two units placed overhead, each centered overa lane on the minor road, each unit consisting of one beacon. When intersection type wasinvestigated, only one group had a significant reduction in crash rate—4-leg intersections with 2-lane main and minor approaches.

Table G-11. Change in Crash Rates When Four-Way Stop Control with Flashing BeaconsAre Added to Intersections with Various Types of Traffic Control (8).

Previous ControlPercent Change

Crash Type Severity

SingleVehicle

MultipleVehicle

PropertyDamage Injury Fatal

2-Way Stop4-Way StopRed-Yellow Flashers

- 30- 100- 10

- 71- 7- 87

- 5770-76

-71- 65 -95

- 100- 100- 100

The characteristics of traffic flow at rural, low-volume intersections controlled by stop signs andby intersection control beacons in conjunction with stop signs were examined (29). The studyfound that intersection control beacons generally reduced vehicular speeds in the majordirections, particularly at intersections with inadequate sight distance. The intersection controlbeacons had, in general, little or no impact on accepted or rejected gaps. A large proportion ofdrivers (40 to 90 percent) violated stop sign laws by not completely stopping at the intersections. The intersection control beacons were not necessarily effective in reducing stop sign violations orcrashes. Guidelines for installation of intersection control beacons are included in the report.


G-15

ILLUMINATION

The objective of a fixed lighting system is to supplement the headlights of automobiles and torender objects that are distant, complex, or that have low contrast more visible to motorists andpedestrians. Specific values on crash experiences related to lighting are not available. Becauseof costs, continuous lighting systems are not generally employed in rural areas; however, lightingsystems can improve safety at isolated, rural at-grade intersections.

The reader should consult both the Roadway Lighting Handbook (30) and the Addendum toChapter Six of the Roadway Lighting Handbook — Designing the Lighting System UsingPavement Luminance (31) for more information on the design of a lighting system.

Lighting should be considered at a rural intersection if the average number of nighttime crashes(N) per year exceeds the average number of day crashes (D) per year divided by 3. If the N isgreater than D/3, the likely average benefit should be taken as N-D/3 crash per year. Abenefit/cost analysis should then be performed to determine if the benefits of lighting theintersection exceed the cost of providing the lighting system (32).

Public lighting of roads is widely accepted as an effective road crash countermeasure. Numerousstudies determined the effects of public lighting on the number of crashes. A synthesis of safetyresearch related to traffic control and roadway elements summarized the results of research andfound that “night crashes can be substantially reduced in number and severity by the use of goodroad lighting” (33). A quantitative meta-analysis of 37 evaluation studies was conducted todetermine the safety effects of public lighting and to examine the validity of the combined results(34). The results of the evaluation studies were the same for all three environments: urban, rural,and freeway. In addition, roadway lighting appears to have a greater effect on pedestrian crashesthan on other types of crashes and a greater effect at junctions than at other locations. It wasconcluded that the best estimates of the safety effects of public lighting are, in rounded values, a65 percent reduction in nighttime fatal crashes, a 30 percent reduction in nighttime injurycrashes, and a 15 percent reduction in nighttime property-damage-only crashes.

PAVEMENT MARKINGS

Pavement markings are used to supplement the regulations or warnings of other devices such astraffic signs or signals (21). They are also used alone to produce results that cannot be obtainedwith other types of traffic control devices. In such cases, they serve as a very effective means ofconveying certain regulations and warnings that could not otherwise be made clearlyunderstandable. The MUTCD (21) provides information on the use and installation of pavementmarkings along roadways and at intersections.

At intersections, pavement markings can be used to help guide vehicles through the intersectionor through the turns. White dotted lines are typically used when guiding vehicles through a turnwithin the shared intersection area. Solid lines are used along the approaches of the intersection.


G-16

In some cases, wider lines or raised pavement markings are used. A stop line is used to stopvehicles in advance of crosswalks or areas where pedestrians are crossing or to indicate where avehicle is to stop at a stop sign or traffic signal. Crosswalks are used as a guide to pedestriansand as a warning to motorists of a pedestrian crossing point. Words and symbol messages on thepavement are also used to guide, warn, or regulate traffic.

SIGHT DISTANCE

Clear sight distance areas should be established, where possible, to ensure that obstructions donot infringe on the sight lines needed by motorists, pedestrians, or bicyclists approachingpotential conflict points. The AASHTO Green Book (5) includes detailed descriptions of how todetermine sight distance along the approaches at an intersection and across their intersectingcorners for a distance sufficient to allow motorists, approaching simultaneously, to see each otherin time to accelerate, slow down, or stop before a collision occurs. In addition, the Green Bookdiscusses the process to determine the necessary sight distance for a driver to make a safedeparture through the intersection area from a stop position.

Table G-12 lists suggested countermeasures for intersections with sight distance concerns.

Sight distances at five intersections were improved in a before-and-after study in Concord,California. Total crashes at these intersections dropped from 39 in the year before to 13 in theyear after obstruction removal (67-percent reduction). In the same study, many otherintersections at other locations in Concord were improved by use of signal installation ormodification, delineation striping, improved pavement markings, and increased policeenforcement. Although all improvements resulted in a reduction in crashes, the greatestpercentage of reduction was experienced at the intersections where the sight distances wereimproved (35).

The IHSDM crash prediction algorithm for rural two-lane highways incorporates the judgment byan expert panel on sight distance obstructions. At a stop control on minor road intersection, whena sight distance obstruction results in a difference of 12.4 mph (20 km/h) or more between thecalculated speed for the available sight distance and design speed, the crash frequency wouldincrease by approximately 5 percent (37).


G-17

Table G-12. Suggested Countermeasures for Intersections (36).

Rural Uncontrolled Intersections• Enact maximum statewide (or county-wide) speed limit of 45 to 50 mph (72.5 to 80.5 km/h) for unsigned

roads.• Formulate simple agreements with property owners to provide obstruction-free corner triangles as large as

possible.• Cut back vegetation and/or embankments to achieve the Green Book sight distance values.• Remove walls, fences, signs, or other obstructions on right-of-way.• Use 2-ft (0.6 m) object height where possible for nighttime view of headlights.• Use speed zoning on approach to intersection.• Place two-way stop signs where Green Book sight distance values cannot be obtained in all four quadrants.

Stop-Controlled Intersections• Cut back vegetation and/or embankments as far as possible.• Restrict parking.• Paint stop line closer to the intersection when that position offers a clearer sight line, and install sign stating

“Pull Up To See.”• Install traffic signals when warranted.• Remove walls, fences, advertising signs, or other obstructions on right-of-way.• Reduce through roadway approach speed limits.• Install four-way stop signs when adequate sight distance cannot be achieved.• Use 2-ft (0.6 m) object height where possible for nighttime view of headlights.


G-18

REFERENCES

1. Hummer, J., C. Hultgren, A. Khattak, T. Hao, N. Stamatiadis, S. Jones, L. Aultman-Hall,and M. Hill. Accidents on Secondary Highways and Countermeasures. Final Report. STCYear 9. Southeastern Transportation Center, Knoxville, TN (April 1998).

2. Foody, T. J., and W. C. Richardson. Evaluation of Left-Turn Lanes as a Traffic ControlDevice. Ohio DOT (1973).

3. Agent, K. R., and R. C. Deen. “Relationship Between Roadway Geometrics andAccidents.” Transportation Research Record 541. Transportation Research Board.Washington, D.C. (1975).

4. Hanna, J. T., T. E. Flynn, and L. T. Webb. “Characteristics of Intersection Accidents inRural Municipalities,” Transportation Research Record 601, Transportation Research Board, Washington, D.C. (1976).


6. Pline, J. “Left-Turn Treatments at Intersections.” Synthesis of Highway Practice Report225, Transportation Research Board, Washington, D.C. (1996).

7. Neuman, T. R. “Intersection Channelization Design Guide.” NCHRP Report 279.Transportation Research Board, Washington, D.C. (1985).

8. Evaluation of Minor Improvements: Flashing Beacons, Safety Lighting, Left-TurnChannelization. Traffic Department, California Department of Public Works (1967).


10. Traffic Engineering Handbook 4th Edition. Institute of Transportation Engineers. NewYork (1999).

11. Sebastian, O. L., and R. S. Pusey. “Paved-Shoulder Left-Turn By-Pass Lanes: A Report onthe Delaware Experience.” Delaware DOT (October 1982).

12. Buehler, M. G. “By-Pass Lanes at Intersections on Two-Lane Roads,” Presented at Thirty-First Annual Illinois Traffic Engineering Conference, University of Illinois, (October1978).


G-19

13. Stover, V. G., W. G. Adkins, and J. C. Goodknight. NCHRP Report 93: Guidelines of Medial and Marginal Access Control on Major Roadways. Highway Research Board,Washington, D.C. (1970).

14. Glennon, J. C., J. J. Valenta, B. A. Thorson, and J. A. Azzeh. Technical Guidelines for theControl of Direct Access to Arterial Highways, Vol. II: Detailed Description of AccessControl Techniques. Report FHWA-RD-76-87. FHWA, U.S. DOT (August 1975).

15. Cottrell, B. H., Jr. The Development of Criteria for the Treatment of Right-Turn Movementson Rural Roads. Final Report VHTRC 81-R45. Virginia Highway and TransportationResearch Council, Charlottesville, VA (March 1981).

16. McCoy, P. T., J. A. Bonneson, S. Ataullah, and D. S. Eitel. “Guidelines for Right-TurnLanes on Urban Roadways.” Transportation Research Record 1445, TransportationResearch Board, Washington, D.C. (1994).

17. Hasan, T., and R. W. Stokes. “Guidelines for Right-Turn Treatments at UnsignalizedIntersections and Driveways on Rural Highways.” Transportation Research Record 1579,Transportation Research Board, Washington, D.C. (1997).

18. Stover, V. G., and F. J. Koepke. Transportation and Land Development. Institute ofTransportation Engineers. (1988).

19. Dale, C. W. Safety Improvement Projects: Summary. Federal Highway Administration.Washington, D.C. (February 1971).

20. Strate, H. E. “Making Highways Safe: A Realistic Approach.” ITE Journal. Institute ofTransportation Engineers. (March 1980).

21. Texas Manual on Uniform Traffic Control Devices for Streets and Highways. TexasDepartment of Transportation. 2003. http://www.dot.state.tx.us/TRF/mutcd.htm.

22. Lyles, R. W. “Evaluation of Signs for Hazardous Rural Intersections.” TransportationResearch Record 782. Transportation Research Board, Washington, D.C. (1980), pp. 22-30.

23. Agent, K. R. “Traffic Control and Accidents at Rural High-Speed Intersections.” Transportation Research Record 1160, Transportation Research Board, Washington, D.C. (1988), pp. 14-21.

24. Roy Jorgensen and Associates. “Cost Effectiveness of Highway Design Elements.” NCHRP Report 197. Transportation Research Board, Washington, D.C. (1978).

25. Gattis, J. L. “Cross Traffic Signing for Stop Signs.” Transportation Research Record 1553.Transportation Research Board, Washington, D.C. (1996), pp. 1-11.

http://www.dot.state.tx.us/TRF/mutcd.htm


G-20

26. Picha, D. L., C. E. Schuckel, J. A. Parham, and C. T. Mai. Traffic Control Devices at Two-Way Stop-Controlled Intersections. FHWA/TX-97/1374-1F. (November 1996).

27. Cribbins, P. D., and C. M. Walton. “Traffic Signals and Overhead Flashers at RuralIntersections: Their Effectiveness in Reducing Accidents,” Highway Research Record 325,Highway Research Board, Washington, D.C. (1970).

28. Foody, T. J., and W. C. Taylor. “An Analysis of Flashing Systems.” Highway ResearchRecord 221, Highway Research Board, Washington, D.C. (1968), pp. 72-84.

29. Pant, P. D., Y. Park, and S. V. Neti. Development of Guidelines for Installation ofIntersection Control Beacons. Final Report. FHWA/OH-93/006. (1992).

30. Roadway Lighting Handbook. Report No. FHWA-IP-78-15. FHWA, U.S. Department ofTransportation, Washington, D.C. (1978).

31. Keck, M. E. Addendum to Chapter Six of the Roadway Lighting Handbook — Designingthe Lighting System Using Pavement Luminance. Report No. FHWA-IP-78-15 (1983Addendum). FHWA, U.S. DOT, Washington, D.C. (September 1983).

32. Wortman, R.H., M. E. Lipinski, L. B. Frickle, W.P. Grimwade, and A.F. Kyle. Development of Warrants for Rural At-Grade Intersection Illumination. University ofIllinois. (1972).

33. Schwab, R. N., N. E. Walton, J. M. Mounce, and M. J. Rosenbaum. Roadway Lighting. Synthesis of Safety Research Related to Traffic Control and Roadway Elements, Vol. 2.,Report FHWA-TS-82-233. FHWA, U.S. DOT, (1982).

34. Elvik, R. “Meta-Analysis of Evaluations of Public Lighting as Accident Countermeasure.”Transportation Research Record 1485, Transportation Research Board, Washington, D.C.(1995).

35. Mitchell, R. “Identifying and Improving High Accident Locations.” Public Works. (December 1972).

36. The Traffic Safety Toolbox: A Primer on Traffic Safety. Institute of TransportationEngineers, (1993).

37. Harwood, D. W., F. M. Council, E. Hauer, W. E. Hughes, and A. Vogt. Prediction of theExpected Safety Performance of Rural Two-Lane Highways, Draft Report No. FHWA/RD-4584-09, Federal Highway Administration, McLean, VA (November 1998).

APPENDIX H: WILDLIFE CRASHES

H-1

Transportation routes can have an effect on animals throughout North America. Although theliterature varies with regard to the amount of displacement and other impacts, there is irrefutableevidence that roads and their associated disturbances reduce habitat effectiveness. This results inreduced fitness and in increased risk of mortality (1).

Crashes between large animals, especially deer, and vehicles are a significant safety problem fora number of rural two-lane highways. According to recent estimates, the number of white taildeer in the lower 48 states (approximately 25 million) has almost doubled in the past decade andis expected to continue increasing in the future (2). The increase in numbers of deer and theirbehavior around highways may explain why deer are involved in so many crashes on ruralhighways. Deer are attracted to highways, partly because of salt leeching into the surroundingsoil, and partly because of forage planted in the median and along the roadway. Additionally,deer cross roadways to move from open feeding areas to protected bedding areas in regularcycles, sometimes several times a day. Deer-vehicle crashes in recent years are estimated to rangefrom 12,000 to 16,000 per year in Minnesota. The average vehicle damage is estimated to be$2000 per crash, and the recreational cost of a deer is estimated to be $500. Therefore, theroadkill of white-tailed deer in Minnesota is about a $35 million problem each year (3).

A review of five states’ crash databases revealed that vehicle-animal crashes increased 69 percentbetween 1985 and 1991. In one state, vehicle-animal crashes composed more than one-third ofall reported vehicle crashes on two-lane rural roads (4). These trends are expected to continue inthe immediate future, increasing the potential for vehicle-animal crashes. The following wereidentified as part of the study:

• The information available indicates that deer are the animal most frequently involved incrashes. In Michigan, almost all reported vehicle-animal crashes were deer related (97.6percent) or deer associated (2.2 percent). Data from Minnesota indicate that deer wereinvolved in more than 90 percent of all reported animal crashes.

• In addition to increasing in frequency, vehicle-animal crashes have increased as a percentageof all reported crashes, from 4.7 percent in 1985 to 8.2 percent in 1991. These figuresindicate that vehicle-animal crashes are increasing at a rate that substantially exceeds that ofother types of crashes. This increase could be a result of continued development andchanging land use patterns, increases in deer population, and increases in traffic volumethrough areas populated by deer.

• Vehicle-animal crashes also occurred more frequently at night. Of all reported animalcrashes, 69 percent to 85 percent occurred at night. The average annual animal crashfrequencies were found to be two to five times higher at night than during the day. Thegreatest number of animal crashes occurred during the early morning hours (5 to 8 am) andthe night hours (6 pm to midnight).

• The greatest number of reported vehicle-animal crashes occurred in November, with thesecond highest in October. These months represent mating season for the deer.


H-2

Several studies on various types of wildlife crashes are summarized on the following pages.

DEER

Countermeasures used to decrease deer crashes have included signing, improvements to roadsidevegetation, reflectors designed to redirect the light from vehicle headlights into the neighboringterrain, fences and underpasses, and a highway crosswalk system.

Signing

When an area is known to have significant deer activity or a deer-vehicle crash history, anadvance deer crossing warning sign may be installed. The MUTCD states that “advance crossingsigns should be used to alert vehicle operators to unexpected entries into the roadway bypedestrians, trucks, bicyclists, animals, and other potential conflicts” (5). The effectiveness ofadvance warning signs is unknown. A concern with their use is that overuse of the deer crossingwarning signs may result in a lack of attention to the message on the part of the motorists.

Roadside Vegetation

One method used to minimize or control the movement of deer onto the roadway is throughimprovements to roadside vegetation and landscape management. A 1980s study in Utah foundthat some deer collisions can be avoided by placing food at points away from the highways (6). These feeding areas intercepted foraging deer and kept them away from the highways, makingthe roads safer for passing motorists. A high big-game fence was installed along an interstate toforce deer to use specified locations for passing under the freeway. The passes were baited withalfalfa hay, fresh vegetable trimmings, and apple pulp to help lure the deer to the underpass. Difficulties associated with the fences included selection of the proper area for the fence, inadequacy of deer guards on ramps of an interchange, and the need for continuous monitoringfor holes in the fence (7).

Reflective Devices

Reflective devices have been designed to redirect the light from vehicle headlights to create“optical fences” to keep deer from crossing or entering the roadway. The reflectors are installedon both sides of the road and reflect the headlights as red lights into the adjoining terrain. Thetheory is that white-tailed deer are afraid of the illuminated red reflectors to the point that theyeither stop or run away when the reflectors are illuminated.

Studies have attempted to determine whether the red reflective devices are an effective treatmentfor reducing deer-vehicle crashes. Some studies have shown that fewer deer are hit when thereflectors are used; however, other studies have demonstrated that the deer are not reacting to thered reflectors as anticipated (8, 9).

In 1980, the Minnesota Department of Transportation (Mn/DOT) installed the Swareflex brand red reflector along a one-mile stretch of I-94 in central Minnesota. Another brand of white


H-3

reflector was installed on a one-mile stretch of TH 169 in the Minnesota River valley in southernMinnesota. The red reflector reduced deer-vehicle crash rates over 80 percent while the whitereflector was unsuccessful. Minnesota has since installed reflectors at 38 locations throughoutthe state. Later installations indicated that reflector installations apparently work in ruralMinnesota and failed in suburban areas. The theory for the success of reflector installations isthat headlights of approaching vehicles shine into reflectors located parallel to the roadway andthe prisms reflect a red glow visible to deer on the roadside. This red glow, perhaps mimickingthe eyes of predators, causes deer to remain motionless or escape away from the roadway whilevehicles are present. The necessity for headlights means they will function as intended onlyduring nighttime and other low light conditions. Deer are most active and deer vehicle crashesoccur predominantly during night or low light conditions. High traffic, increasing deerpopulation, and the inability to effectively maintain the reflectors may have also been factors inthe lack of success in the metropolitan area (3).

Fences and Underpasses

Use of fencing and underpasses has resulted in fewer deer crossing the roadways and fewercrashes (7, 10). A study of two segments of 8-ft (2.4 m) fences with one-way gates in Minnesotafound that the reported number of deer hits was reduced 60 and 93 percent from the expectednumber for the two segments (10).

Highway Crosswalk System

A new mitigative technique was studied in Summit and Wasatch counties in northeastern Utah. Analysis of designated kill zones compared to non-kill zones on each highway helped identifydistinguishing features that aided placement of the crossing structures. The percent of vegetativecover was higher for designated kill zones (40 percent) compared to non-kill zones (29 percent)(11).

The crosswalk system restricted deer crossings to specific, well-marked areas along the highwayswhere motorists could anticipate them. Right-of-ways were fenced off with deer-proof fencing todirect the animals to the designated crossing areas. At these locations, deer jumped a 3.3-ft (1.0 m) high fence to enter the crosswalk funnel. Once in the funnel, the animal could choose toforage on desired vegetation or continue to approach the road. The fence could not extend closerthan 30 ft (9.1 m) from the highway surface, so fields of rounded river cobbles were used todemarcate a path for the deer to follow as it continued to approach the road. Painted cattle-guardlines on the road surface were used to delineate crosswalk boundaries for oncoming motoristsand may have served as a visual cue to guide deer directly across the highway. Once across theroad, the deer encountered another dirt path bordered by cobbles, and a narrow fence openingallowing entry to the crosswalk funnel and distant habitat. This system is illustrated in Figure H-1.

Vegetation in and along cobble paths was eliminated to discourage deer from remaining near thehighway. A series of three warning signs was installed at each crosswalk to advise motorists thatthey were entering a crossing zone. Four one-way grates were installed in the vicinity of each


H-4

Figure H-1. Major Crosswalk Features on (a) Two-Lane Highways and (b) Four-Lane Highways (11).


H-5

crosswalk to enable deer that became trapped along the highway corridor to escape to the right-of-way.

This study represents the initial implementation and testing of the crosswalk system. Thecrosswalks were used because they could be easily installed along the existing roadways at one-sixth the cost required to excavate tunnels and install underpasses. Observations of deersuccessfully crossing within crosswalk boundaries, the apparent maintenance of migratorybehavior, and reduced deer use of the highway right-of-way indicate the system warrants furthertesting. This study also identified problems in the original design so that modification can bemade (11).

LARGE MAMMALS

The Bow River Valley is rich in natural, wildland resources, particularly wildlife. Banff NationalPark has 54 species of mammals and 280 species of birds. The TransCanada Highway (TCH)has directly impacted many of these species of wildlife or affected their habitats. Originalconstruction in the 1950s had realigned the river in numerous locations. When planning toupgrade the roadway, it was clear that environmental protection would be a major objective and ascientific challenge (12).

Elk, mule and white-tailed deer, moose, black bears, coyotes, bighorn sheep, and smallermammals, such as pine squirrels and hare, were regularly killed on the highway. Occasionally,grizzly bear, wolf, wolverine, lynx, marten, porcupine, hawk, owl, and others were struck (12). Banff National Park and Alberta Provincial records (Alberta, Canada) have documented thenumber of carnivores killed in vehicle collision in the past 10 years (see Table H-1). This mustbe considered a minimum as the animals that were hit but never found have not been recorded.

It was decided to fence both sides of the new roadway with an 8-ft (2.4 m) high page wire fence.A system of 10 wildlife underpasses was also installed in an effort to mitigate wildlife crashes.Texas gates and stiles were used to allow unimpeded vehicular and pedestrian passage throughthe fences. One-way and conventional gates were installed for wildlife management actions. Fish habitat was recreated where major fish-bearing streams were impacted. Underpasses variedfrom conventional, bridge-like concrete structures with 42.8-ft (13 m) span openings and 13.2-ft(4 m) headway to 13.2-ft (4 m) circular culverts and 13.2 ft (4 m) by 23.0 ft (7 m) ellipticalmultiplate culverts. Underpasses varied depending on the centerline to centerline separation ofthe roadway. By 1990, 19.3 mi (31 km) of twinned highway and 10 underpasses had beenconstructed.

The research revealed the fences to be highly effective in reducing wildlife collisions—over 94percent for elk. Other large species were similar. Detailed research of deer has not been pursuedalthough tracking beds show that deer use the underpasses. Most other highly transient species,such as wolf, grizzly and black bear, bighorn sheep, coyote, lynx, and some small mammals,have been recorded using the underpasses (12). However, problems have been identified andseveral unexpected wildlife impact occurrences were recorded.


H-6

Table H-1. Highway and Railway Mortality of Large Carnivores in the Bow River Valley, Alberta, Canada (12).

Species Inside Banff National Park Outside Banff National Park Total

Highway Rail Highway Rail

Coyote 117 7 39 1 164

Black bear 12 5 8 2 27

Cougar 1 0 2 0 3

Grizzly bear 1 0 0 0 1

Wolverine 2 0 0 0 2

Lynx 0 0 4 0 4

TORTOISES AND SMALL VERTEBRATES

Roads and highways impact tortoise populations through restriction of movement in addition todirect mortality and facilitating illegal collections. Because there are many roads and highwaysthrough the habitat of the desert tortoise, the potential for road kills to affect tortoise populationsis high. Consequently, reducing road kills could help to facilitate recovery of tortoisepopulations. Barrier fences are a potential mitigation measure, but they also increase populationfragmentation. Culverts beneath the roadway may reduce fragmentation by facilitatingmovements of tortoises between both sides of the road (13).

A scientific research project was designed to learn the effectiveness of a highway barrier fencebuilt to aid in the recovery of desert tortoise population along California State Highway 58 (Hwy58) in the western Mojave Desert of California. In 1990, the California Department ofTransportation (CalTrans) erected tortoise-barrier fencing along a section of Hwy 58 that wasscheduled for widening from two lanes to four lanes. Several agencies joined the cooperativemonitoring project to learn the effectiveness of protective fencing and culverts. The 14.9 mi (24 km) long fence consists of 2-ft (60 cm) wide, 0.5-ft (1.3 cm) mesh, galvanized steel,hardware cloth that is buried to 5.9 inches (15 cm) beneath ground level and extends 17.7 inches(45 cm) above the ground. The fence is supported by a six-strand wire fence; the top threestrands are barbed to inhibit access by humans and livestock, and the three bottom strands areunbarbed to allow easy installation of the hardware cloth and to allow medium-sized mammals toclimb over without being injured. The bottom two strands are placed beneath the top of thehardware cloth to provide structural support to the cloth. The wires are attached to the cloth bysteel rings. The fence is held up by 6.6-ft (2 m) t-bars spaced approximately 9.9 ft (3 m) apart. Gates, which are required to allow access to private property along the highway edge, were alsodesigned as barriers to tortoises. The same hardware cloth that is used on the fence is separatelyattached to the lower part of the gate. The gates are hung close to the ground and flush to 7.9inches (20 cm) by 7.9 inches (20 cm) wood beams buried between gate posts to prevent tortoisesfrom escaping under the gates.


H-7

Twenty-four culverts that span the entire width of the highway are in place and are all designedfor rainwater runoff. In August 1992, the fence on Hwy 58 was attached in funnel fashion tostorm-drain culverts to facilitate movements by tortoises under the highway. The culverts aremade of 3 ft (0.9 m) to 4.9 ft (1.5 m) diameter corrugated steel pipe, 4.6 ft (1.4 m) diameterreinforced concrete pipe, or 9.9 ft (3 m) to 11.8 ft (3.6 m) by 5.9 ft (1.8 m) to 9.9 ft (3 m)reinforced concrete boxes. The culverts are 109 ft (33 m) to 217 ft (66 m) long.

Researchers conducted surveys in July of 1992, 1993, and 1994 and recorded the identity andlocation of all animal carcasses. A total of 1080 carcasses were found, including 36 tortoisecarcasses. Researchers searched for the carcasses along a 14.9-mi (24 km) section of fencedhighway and along a 14.9-mi (24 km) section of unfenced highway. They found 88 percentfewer vertebrate carcasses and 93 percent fewer tortoise carcasses along the fenced section ofhighway. These differences were highly significant and indicate that the fence was verysuccessful at reducing road mortality. However, in 1995, several tortoises were killed along thefenced section of Hwy 58, all within 0.3-mi (0.5 km) gaps in the fence. Most of the gaps weredue to poor maintenance, indicating that proper maintenance of the fence is critical to its success.

The results indicate that, when new or properly maintained, the barrier fence was effective ingreatly reducing highway mortality in several species of vertebrates, including the threatenedtortoise. However, tortoises can escape from relatively small gaps in improperly installed ormaintained fences and gates. Tortoises and other vertebrates also used culverts, but it has not yetbeen determined whether their use will reduce the fragmenting effects of the fence and highway. Culvert use is expected to increase with time as more animals settle near and discover theculverts (13).

BROWN PELICANS

The Texas Department of Transportation (TxDOT) is continuing efforts to eliminate theaccidental deaths and injuries of endangered brown pelicans on the Queen Isabella Causeway. The Queen Isabella Causeway is a 2.4-mi (3.9 km) long, four-lane bridge connecting Port Isabeland South Padre Island. The bridge center span rises 84 ft (25.6 m) above the Gulf IntercostalWaterway (14).

The eastern brown pelican is a large bird with an average weight of 7.5 lb (2.8 kg), a body lengthof 4 ft (1.2 m), and a wingspan of 6.5 ft (2 m). It flies 14 to 35 mph (12 to 56 km/h), often withslow wing beats close to the water. The brown pelican is a coastal resident that seldom straysinland. These large birds land on the Queen Isabella Causeway and are sometimes struck byvehicles.

The first reported death of a brown pelican on the Causeway was in September 1984. Since then,a number of brown pelican deaths have been documented between September and early Marcheach year. The increasing traffic mortality of the endangered birds prompted a 1988-1990 studyby the Texas Transportation Institute. This study, coupled with wind tunnel studies of theairflow around models of the bridge, led to the conclusion that the mortalities result from acombination of several factors:


H-8

• an increase in pelican population,• flight patterns of the birds as they fly to roosting sites in the evening,• the occasional presence of strong northerly winds and inclement weather, and• air flow patterns above the bridge deck.

The study concluded that the birds are not intentionally landing on the bridge deck. Rather,turbulence above the deck causes the birds to land if they attempt to fly over the bridge withoutsufficient initial altitude. The study especially indicates a connection between pelican deaths andthe passage of cold fronts accompanied by strong wind (northers). The study determined thatflashing lights, propane cannon, or other noise makers are not likely to discourage pelicans fromintentionally landing. Alternate roosting structures and platforms or additional railing on thebridge were not effective. The study identified traffic control measures as the actions most likelyto effectively reduce pelican mortalities.

As a result of meetings with many interested agencies and recommendations from the TTI report,TxDOT took the following actions:

• Flashing signs to reduce speed were installed at each end of the bridge and at the crest of thebridge. These signs were installed after it was determined that a silhouette sign previouslyinstalled was not effective.

• Lights on the causeway were adjusted to come on 30 minutes earlier in the evening.• Changeable messages were installed at each end of the bridge to warn motorists to slow down

and drive cautiously for conditions that may exist on the bridge.• Windsocks and banners to distract the pelicans were installed on light poles at the crest of the

bridge.• A “Pelican Patrol” consisting of TxDOT personnel was established to patrol the bridge

during northers to pick up or assist downed pelican and activate the warning signs.• A plan was established to determine who would pick up the birds and where they would be

taken. These measures are active during northers and inclement weather months, specificallyfrom September through February.

In addition, a public service announcement was produced by TxDOT and has been airing onlocal, national, and international television stations. The announcement was intended to makethe public aware of the pelican population and its endangered status. The announcementencourages motorists to reduce speed on the causeway and provided information on how to assistdowned or injured pelicans. Four pelicans died during the winter following the use of thesemeasures compared to eight during the previous winter. TxDOT is also considering otherpossible mitigation measures including adding more banners to the Causeway, a publicitycampaign to include flyers and posters, adding call boxes at each end of the causeway, andinstallation of weather monitoring devices to detect northers (14).

BATS

Although not directly related to wildlife crashes, TxDOT has initiated a study of bats andbridges. The knowledge developed in the Bats and Bridges Study is helping to define how to


H-9

include bats in a new bridge design where appropriate and to exclude them where not desired(14).

OCELOTS

The ocelot is a medium-sized, spotted and blotched cat with a moderately long tail and is afederally listed endangered species. The cats once ranged over the southern part of Texas withoccasional records from north and central Texas, but they are now restricted to several isolatedpatches of suitable habitat in three or four counties of the Rio Grande Plains.

The major cause of mortality for the ocelot population has been ocelot-automobile collisions. In1993, TxDOT proposed improvements to State Highway 100 in Cameron County, Texas. Due toreported ocelot sightings (transportation-related mortalities) in the area, TxDOT worked incooperation with wildlife agencies regarding the concern for the ocelot population.

A 48-in (122 cm) pipe culvert in a drainage ditch containing suitable habitat for the ocelot wasinstalled adjacent to an 8 ft (2.4 m) by 5 ft (1.5 m) box culvert and was placed above the usualplane of high water. A 1 ft (0.3 m) wide concrete cat ramp at each end of the culvert was builtfrom the entrance to the edge of the ditch below the level of the berm. Brush was allowed torevegetate the area immediately adjacent to the rip-rap, and a no mow area was established oneither side of the culvert. Finally, a hog-wire fence was constructed after highway constructionwas completed.

TxDOT has installed several ocelot crossings throughout the southern portion of the state. Research is being proposed to ascertain the efficiency of the structures (14).

FLORIDA PANTHERS AND OTHER WILDLIFE IN SOUTHWEST FLORIDA

A contiguous system of wild lands is necessary to accommodate the spatial needs of the Floridapanther population. Adult male and female panthers maintain home ranges of >193 sq mi (500 sqkm) and >73 sq mi (190 sq km), respectively, with limited overlap among males. These homeranges often include many miles of improved roads that are regularly traversed. Road-killmortality can be expected among panthers as a result of the interspersion of roads with pantherhabitat (15).

Efforts to reduce this unnatural source of mortality have included the creation of nighttime speedreduction zones, installation of special roadside headlight reflections, and adding rumble strips tothe highway surface. However, a more ambitious project was completed when State Road 84was converted to Interstate 75.

Locations of previous road-kills and knowledge of where radio-instrumented panthers crossedthis busy highway were used to incorporate 24 wildlife underpasses into the highway conversiondesign. These strategically placed structures offer safe passage to wildlife that is beneath theflow of traffic. Use of these underpasses was encouraged by erecting an 11.2-ft (3.4 m) chain-link fence topped with three strands of outrigged barbed wire along the 40.4-mi (65 km) stretch


H-10

of interstate that runs through panther habitat. A second wildlife crossing design was developedfor State Road (SR) 29, a two-lane highway running through panther habitat. The crossings onSR 29 consisted of a pre-formed box culvert 7.9 ft (2.4 m) high, 24 ft (7.3 m) wide, and 48 ft(14.6 m) long. The culverts rest at ground level, and the roadway gradually rises over theculverts. The crossing also includes a concrete span that forms a bridge across the adjacentcanal. The surface of the span contains a layer of soil to support growth of natural vegetation.This crossing was installed at two critical areas. The SR 29 corridor with the installed crossingswas fenced similarly to I-75.

The objectives were to evaluate the effectiveness of the underpass design installed on SR 29 andto compare the use to the I-75 wildlife crossings. Both designs of wildlife crossings have beenused by Florida panthers and a host of other animal species. The I-75 wildlife crossings withtheir openness and creation of early successional habitat may have encouraged use by white-tailed deer. The more shaded, cooler, and damper SR 29 structures may have created idealhabitat for raccoon prey items, accounting for the heavy use by these mammals. It appears thateither wildlife crossing design will be successful when placed at sites where animals habituallycross (15).

REVIEW OF VEHICLE-ANIMAL CRASHES

The recent review of vehicle-animal crashes developed the following recommendations foradditional consideration (4):

• Policies for installing deer warning signs should be reviewed to limit their use to locationswith significant deer crash problems or areas with high deer activity. In this way, the signsmay become more meaningful for alerting drivers to potentially dangerous situations.

• Warning reflectors should be further evaluated as a low-cost countermeasure.• Other, more sophisticated roadway- or vehicle-based detection devices should be considered

in the development and operational testing of rural intelligent transportation systemapplications.

• Driver education classes in areas with high vehicle-animal crashes should includeinformation on the patterns of animal crashes. If drivers are conditioned to expect a higherchance of encountering deer during November or early morning and early nighttime, theymay be better prepared to react to such a sudden encounter.


H-11

REFERENCES

1. Gibeau, M. L., and K. Heuer. “Effects of Transportation Corridors on Large Carnivores inthe Bow River Valley, Alberta.” Transportation and Wildlife: Reducing Wildlife Mortalityand Improving Wildlife Passageways across Transportation Corridors. Proceedings of theFlorida Department of Transportation/Federal Highway Administration Transportation-Related Wildlife Mortality Seminar. Orlando, Florida. (April 30-May 2, 1996).

2. Nelson, P. “Deer Watch.” National Wildlife. National Wildlife Federation. Washington, D.C.Vol. 32, No. 6 (October-November, 1994), pp. 34-41.

3. Pafko, F., and V. Kovach. “Minnesota Experience with Deer Reflectors.” Transportationand Wildlife: Reducing Wildlife Mortality and Improving Wildlife Passageways acrossTransportation Corridors. Proceedings of the Florida Department of Transportation/FederalHighway Administration Transportation-Related Wildlife Mortality Seminar, Orlando,Florida (April 30-May 2, 1996).

4. Hughes, W. E., A. R. Saremi, and J. F. Paniati. “Vehicle-Animal Crashes: An IncreasingSafety Problem.” ITE Journal (August 1996).

5. FHWA. Manual on Uniform Traffic Control Devices. Federal Highway Administration,Washington, D.C. (2001).

6. 1985-1986 Study on Deer Collisions. Insurance Institute for Highway Safety, Arlington, VA. (1986).

7. Ward, A. L. “Mule-Deer Behavior in Relation to Fencing and Underpasses on Interstate 80 inWyoming.” Transportation Research Record 859. Transportation Research Board,Washington, D.C. (1982).

8. Schafer, J. A., S. Penland, and W. P. Carr. “Effectiveness of Wildlife Warning Reflectors inReducing Deer-Vehicle Accidents in Washington State.” Transportation Research Record1010. Transportation Research Board, Washington, D.C. (1985).

9. Zacks, J. L. “Do White-Tailed Deer Avoid Red? An Evaluation of the Premise Underlyingthe Design of Swareflex Wildlife Reflectors (Discussion and Closure).” TransportationResearch Record 1075. Transportation Research Board, Washington, D.C. (1986).

10. Ludwig, J., and T. Bremicker. “Evaluation of 2.4-m Fences and One-Way Gates forReducing Deer-Vehicle Collisions in Minnesota.” Transportation Research Record 913. Transportation Research Board, Washington, D.C. (1983).

11. Lehnert, M. E., L. A. Romin, and J. A. Bossonette. “Mule Deer-Highway Mortality inNortheastern Utah: Causes, Patterns, and New Mitigative Technique.” Transportation andWildlife: Reducing Wildlife Mortality and Improving Wildlife Passageways across


H-12

Transportation Corridors. Proceedings of the Florida Department of Transportation/FederalHighway Administration Transportation-Related Wildlife Mortality Seminar. Orlando,Florida. (April 30-May 2, 1996).

12. Leeson, B. F. “Highway Conflicts and Resolutions in Banff National Park, Alberta, Canada.” Transportation and Wildlife: Reducing Wildlife Mortality and Improving WildlifePassageways across Transportation Corridors. Proceedings of the Florida Department ofTransportation/Federal Highway Administration Transportation-Related Wildlife MortalitySeminar. Orlando, Florida. (April 30-May 2, 1996).

13. Boarman, W. I., and M. Sazaki. “Highway Mortality in Desert Tortoises and SmallVertebrates: Success of Barrier Fences and Culverts.” Transportation and Wildlife: ReducingWildlife Mortality and Improving Wildlife Passageways across Transportation Corridors. Proceedings of the Florida Department of Transportation/Federal Highway AdministrationTransportation-Related Wildlife Mortality Seminar. Orlando, Florida. (April 30-May 2,1996).

14. Jenkins, K. “Texas Department of Transportation Wildlife Activities.” Transportation andWildlife: Reducing Wildlife Mortality and Improving Wildlife Passageways acrossTransportation Corridors. Proceedings of the Florida Department of Transportation/FederalHighway Administration Transportation-Related Wildlife Mortality Seminar. Orlando,Florida. (April 30-May 2, 1996).

15. Land, D., and M. Lotz. “Wildlife Crossing Designs and Use by Florida Panthers and OtherWildlife in Southwest Florida.” Transportation and Wildlife: Reducing Wildlife Mortalityand Improving Wildlife Passageways across Transportation Corridors. Proceedings of theFlorida Department of Transportation/Federal Highway Administration Transportation-Related Wildlife Mortality Seminar. Orlando, Florida. (April 30-May 2, 1996).

APPENDIX I: WORK ZONES

I-1

Several sources of information are available on work zones including the National Work ZoneSafety Information Clearinghouse (1). Opened in February 1998, the clearinghouse is acooperative venture between the Federal Highway Administration and ARTBA to improve safetyat roadway construction sites. The clearinghouse is operated and maintained by the TexasTransportation Institute and includes an interactive Internet website, on-site research personnel,and customer service representatives. Besides FHWA, ARTBA, and TTI, the clearinghouse isbeing sponsored by the American Association of State Highway and Transportation Officials, theLaborers’ International Union of North America, CNA Commercial Insurance, the InternationalMunicipal Signal Association, the National Association of County Engineers, and Lanford Bros.Company. Marketing partners include the National Utility Contractors Association and theInstitute of Traffic Engineers. The URL is http://wzsafety.tamu.edu/ and the e-mail [email protected].

While the clearinghouse receives a number of phone and fax requests for information, it is thewebsite that services many requests – now receiving over 5000 hits a month. The site includesfive searchable databases on key contact personnel, safety practices, available technologies,research results, and safety training courses and programs. Besides links to other related sites, thesite now offers many materials and even some full reports online.

A TxDOT study produced a catalog of effective treatments to improve driver and worker safetyat short-term work zones on rural highways (2). The catalog provides a brief description of eachtreatment, along with a summary of the treatment’s effectiveness, and recommendations for itsuse at short-term work zones. Devices that were found to be effective included:

• fluorescent yellow-green worker vests and hard hat covers,• portable variable message signs,• speed display trailers,• fluorescent orange roll-up signs,• radar drones, and• retroreflective magnetic strips for worker vehicles.

http://wzsafety.tamu.edu/

APPENDIX I: WORK ZONES

I-2

REFERENCES

1. “Clearinghouse Continues Its Successful Climb to Prominence.” Texas TransportationResearcher, Volume 35, Number 1 (1999).

2. Fontaine, M. D., and H. G. Hawkins, Jr. Catalog of Effective Treatments to Improve Driverand Worker Safety at Short-Term Work Zones. FHWA/TX-01/1879-3 (2001).

APPENDIX J: CRASHES BY AREA/VOLUME

J-1

CRASH FREQUENCIES AND DISTRIBUTIONS FOR ON-SYSTEM, RURAL TWO-LANE HIGHWAYS BY AREA TYPE AND VOLUME

Data Source: Traffic Accident Database, Texas Department of Public Safety, 1997 to 1999, Low-Volume (ADT�2000), Rural Two-Lane Highways

Table J-1. Crashes by Accident Severity.Dev, ADT ACCIDENT SEVERITYFrequencyPercentRow PercentCol Percent

Non-incapacitatingInjury

IncapacitatingInjury

Fatal TOTAL

RURAL, ADT<=2K 8930 4348 1344 14,62218.55 9.03 2.79 30.3861.07 29.74 9.1928.49 32.86 37.78

RURAL, ADT=2-6K 8842 4128 1229 14,19918.37 8.58 2.55 29.5062.27 29.07 8.6628.21 31.20 34.55

RURAL, ADT>6K 4413 1834 521 67689.17 3.81 1.08 14.06

65.20 27.10 7.7014.08 13.86 14.65

URBAN, ADT<=2K 393 164 45 6020.82 0.34 0.09 1.25

65.28 27.24 7.481.25 1.24 1.27

URBAN, ADT=2-6K 2328 834 138 33004.84 1.73 0.29 6.86

70.55 25.27 4.187.43 6.30 3.88

URBAN, ADT>6K 6436 1922 280 863813.37 3.99 0.58 17.9574.51 22.25 3.2420.53 14.53 7.87

TOTAL 31,342 13,230 3557 48,12965.12 27.49 7.39 100


J-2

Table J-2. Crashes by Intersection Related.Dev, ADT INTERSECTION RELATEDFrequencyPercentRow PercentCol Percent

IntersectionIntersection

RelatedDriveway

AccessNon

IntersectionTOTAL

RURAL, ADT<=2K 2248 1346 1269 9759 14,6224.67 2.80 2.64 20.28 30.38

15.37 9.21 8.68 66.74 19.47 22.85 20.98 39.60

RURAL, ADT=2-6K 2942 1479 1896 7882 14,1996.11 3.07 3.94 16.38 29.50

20.72 10.42 13.35 55.5125.48 25.11 31.34 31.99

RURAL, ADT>6K 1612 917 1188 3051 67683.35 1.91 2.47 6.34 14.06

23.82 13.55 17.55 45.0813.96 15.57 19.64 12.38

URBAN, ADT<=2K 230 84 58 230 6020.48 0.17 0.12 0.48 1.25

38.21 13.95 9.63 38.211.99 1.43 0.96 0.93

URBAN, ADT=2-6K 1256 548 378 1118 33002.61 1.14 0.79 2.32 6.86

38.06 16.61 11.45 33.8810.88 9.30 6.25 4.54

URBAN, ADT>6K 3259 1516 1261 2602 86386.77 3.15 2.62 5.41 17.95

37.73 17.55 14.60 30.1228.22 25.74 20.84 10.56

TOTAL 11,547 5890 6050 24,642 48,12923.99 12.24 12.57 51.20 100


J-3

Table J-3. Crashes by First Harmful Event.Dev, ADT FIRST HARMFUL EVENTFrequencyPercentRow PercentCol Percent

Other Non-

CollisionOverturned Pedestrian

Other MotorVehicle

in TransitRR Train Parked Car

RURAL, ADT<=2K 122 4209 145 4548 35 720.25 8.75 0.30 9.45 0.07 0.150.83 28.79 0.99 31.10 0.24 0.49

36.53 52.92 17.79 17.68 46.05 21.24RURAL, ADT=2-6K 104 2352 211 7237 24 111

0.22 4.89 0.44 15.04 0.05 0.230.73 16.56 1.49 50.97 0.17 0.78

31.14 29.57 25.89 28.13 31.58 32.74RURAL, ADT>6K 31 615 134 4586 2 81

0.06 1.28 0.28 9.53 0.00 0.170.46 9.09 1.98 67.76 0.03 1.209.28 7.73 16.44 17.83 2.63 23.89

URBAN, ADT<=2K 6 83 7 351 2 10.01 0.17 0.01 0.73 0.00 0.001.00 13.79 1.16 58.31 0.33 0.171.80 1.04 0.86 1.36 2.63 0.29

URBAN, ADT=2-6K 23 294 63 2199 4 310.05 0.61 0.13 4.57 0.01 0.060.70 8.91 1.91 66.64 0.12 0.946.89 3.70 7.73 8.55 5.26 9.14

URBAN, ADT>6K 48 401 255 6805 9 430.10 0.83 0.53 14.14 0.02 0.090.56 4.64 2.95 78.78 0.10 0.50

14.37 5.04 31.29 26.45 11.84 12.68TOTAL 334 7954 815 25,726 76 339

0.69 16.53 1.69 53.45 0.16 0.70


J-4

Table J-3. Crashes by First Harmful Event (continued).Dev, ADT FIRST HARMFUL EVENTFrequencyPercentRow PercentCol Percent

Pedalcyclist Animal Fixed Object Other Object TOTAL

RURAL, ADT<=2K 71 637 4739 44 14,6220.15 1.32 9.85 0.09 30.380.49 4.36 32.41 0.30

18.49 51.21 42.55 36.97 RURAL, ADT=2-6K 77 401 3645 37 14,199

0.16 0.83 7.57 0.08 29.500.54 2.82 25.67 0.26

20.05 32.23 32.73 31.09 RURAL, ADT>6K 53 101 1152 13 6768

0.11 0.21 2.39 0.03 14.060.78 1.49 17.02 0.19

13.80 8.12 10.34 10.92 URBAN, ADT<=2K 8 6 138 0 602

0.02 0.01 0.29 0.00 1.251.33 1.00 22.92 0.00 2.08 0.48 1.24 0.00

URBAN, ADT=2-6K 53 52 574 7 33000.11 0.11 1.19 0.01 6.861.61 1.58 17.39 0.21

13.80 4.18 5.15 5.88 URBAN, ADT>6K 122 47 890 18 8638

0.25 0.10 1.85 0.04 17.951.41 0.54 10.30 0.21

31.77 3.78 7.99 15.13 TOTAL 384 1244 11,138 119 48,129

0.80 2.58 23.14 0.25 100.00


J-5

Table J-4. Crashes by Object Struck.Dev, ADT OBJECT STRUCKFrequencyPercentRow PercentCol Percent

No Code Applicable

VehicleOver-turned

Hole in Road

Vehicle Jack-knifed

Person Fell or

Jumpedfrom

Vehicle

Vehicle HitTrain onParallelTracks

TrainMoving

Forward

RURAL, ADT<=2K 7261 230 8 33 68 1 3415.09 0.48 0.02 0.07 0.14 0.00 0.0749.66 1.57 0.05 0.23 0.47 0.01 0.2323.99 27.28 61.54 27.50 41.98 25.00 45.95

RURAL, ADT=2-6K 8439 278 4 48 42 3 2217.53 0.58 0.01 0.10 0.09 0.01 0.0559.43 1.96 0.03 0.34 0.30 0.02 0.1527.88 32.98 30.77 40.00 25.93 75.00 29.73

RURAL, ADT>6K 4843 145 0 21 14 0 310.06 0.30 0.00 0.04 0.03 0.00 0.0171.56 2.14 0.00 0.31 0.21 0.00 0.0416.00 17.20 0.00 17.50 8.64 0.00 4.05

URBAN, ADT<=2K 379 12 1 2 4 0 20.79 0.02 0.00 0.00 0.01 0.00 0.00

62.96 1.99 0.17 0.33 0.66 0.00 0.331.25 1.42 7.69 1.67 2.47 0.00 2.70

URBAN, ADT=2-6K 2361 52 0 4 13 0 44.91 0.11 0.00 0.01 0.03 0.00 0.01

71.55 1.58 0.00 0.12 0.39 0.00 0.127.80 6.17 0.00 3.33 8.02 0.00 5.41

URBAN, ADT>6K 6986 126 0 12 21 0 914.52 0.26 0.00 0.02 0.04 0.00 0.0280.88 1.46 0.00 0.14 0.24 0.00 0.1023.08 14.95 0.00 10.00 12.96 0.00 12.16

TOTAL 30,269 843 13 120 162 4 7462.89 1.75 0.03 0.25 0.34 0.01 0.15


J-6

Table J-4. Crashes by Object Struck (continued).Dev, ADT OBJECT STRUCKFrequencyPercentRow PercentCol Percent

TrainBacking

TrainStanding

Still

Train/Action

Unknown

Highway Sign

CurbCulvert/

HeadwallGuardrail

RURAL, ADT<=2K 0 1 0 546 14 727 1730.00 0.00 0.00 1.13 0.03 1.51 0.360.00 0.01 0.00 3.73 0.10 4.97 1.180.00 33.33 0.00 37.22 15.73 43.20 26.57

RURAL, ADT=2-6K 2 2 0 472 14 583 2410.00 0.00 0.00 0.98 0.03 1.21 0.500.01 0.01 0.00 3.32 0.10 4.11 1.70

100.00 66.67 0.00 32.17 15.73 34.64 37.02RURAL, ADT>6K 0 0 1 172 8 198 114

0.00 0.00 0.00 0.36 0.02 0.41 0.240.00 0.00 0.01 2.54 0.12 2.93 1.680.00 0.00 100.00 11.72 8.99 11.76 17.51

URBAN, ADT<=2K 0 0 0 28 1 14 60.00 0.00 0.00 0.06 0.00 0.03 0.010.00 0.00 0.00 4.65 0.17 2.33 1.000.00 0.00 0.00 1.91 1.12 0.83 0.92

URBAN, ADT=2-6K 0 0 0 91 6 73 390.00 0.00 0.00 0.19 0.01 0.15 0.080.00 0.00 0.00 2.76 0.18 2.21 1.180.00 0.00 0.00 6.20 6.74 4.34 5.99

URBAN, ADT>6K 0 0 0 158 46 88 780.00 0.00 0.00 0.33 0.10 0.18 0.160.00 0.00 0.00 1.83 0.53 1.02 0.900.00 0.00 0.00 10.77 51.69 5.23 11.98

TOTAL 2 3 1 1467 89 1683 6510.00 0.01 0.00 3.05 0.18 3.50 1.35


J-7

Table J-4. Crashes by Object Struck (continued). Dev, ADT OBJECT STRUCKFrequencyPercentRow PercentCol Percent

RR SignalPole

RRCrossing

Gates

SignalPole/Post

SignalLight/Wires

Work ZoneBarricade

LuminairePole

UtilityPole

RURAL,ADT<=2K

8 3 13 0 3 16 3000.02 0.01 0.03 0.00 0.01 0.03 0.620.05 0.02 0.09 0.00 0.02 0.11 2.05

33.33 37.50 12.62 8.33 13.01 28.22RURAL, ADT=2-6K

8 1 14 0 10 19 3080.02 0.00 0.03 0.00 0.02 0.04 0.640.06 0.01 0.10 0.00 0.07 0.13 2.17

33.33 12.50 13.59 27.78 15.45 28.97RURAL, ADT>6K 2 2 5 0 10 17 146

0.00 0.00 0.01 0.00 0.02 0.04 0.300.03 0.03 0.07 0.00 0.15 0.25 2.168.33 25.00 4.85 27.78 13.82 13.73

URBAN,ADT<=2K

0 0 1 0 0 2 240.00 0.00 0.00 0.00 0.00 0.00 0.050.00 0.00 0.17 0.00 0.00 0.33 3.990.00 0.00 0.97 0.00 1.63 2.26

URBAN, ADT=2-6K

4 1 14 0 3 17 850.01 0.00 0.03 0.00 0.01 0.04 0.180.12 0.03 0.42 0.00 0.09 0.52 2.58

16.67 12.50 13.59 8.33 13.82 8.00URBAN, ADT>6K 2 1 56 0 10 52 200

0.00 0.00 0.12 0.00 0.02 0.11 0.420.02 0.01 0.65 0.00 0.12 0.60 2.328.33 12.50 54.37 27.78 42.28 18.81

TOTAL 24 8 103 0 36 123 10630.05 0.02 0.21 0.00 0.07 0.26 2.21


J-8


Mailbox Tree/Shrub FenceHouse/

BuildingCommercial

Sign

OtherFixed

Object

MaintenanceBarricade or

Materials

RURAL,ADT<=2K

165 1411 1972 44 7 338 20.34 2.93 4.10 0.09 0.01 0.70 0.001.13 9.65 13.49 0.30 0.05 2.31 0.01

37.41 46.31 51.37 30.99 14.89 38.11 28.57RURAL,ADT=2-6K

154 1041 1218 44 8 261 10.32 2.16 2.53 0.09 0.02 0.54 0.001.08 7.33 8.58 0.31 0.06 1.84 0.01

34.92 34.16 31.73 30.99 17.02 29.43 14.29RURAL,ADT>6K

70 255 283 14 8 114 30.15 0.53 0.59 0.03 0.02 0.24 0.011.03 3.77 4.18 0.21 0.12 1.68 0.04

15.87 8.37 7.37 9.86 17.02 12.85 42.86URBAN,ADT<=2K

4 34 47 3 0 5 00.01 0.07 0.10 0.01 0.00 0.01 0.000.66 5.65 7.81 0.50 0.00 0.83 0.000.91 1.12 1.22 2.11 0.00 0.56 0.00

URBAN,ADT=2-6K

19 150 140 14 9 58 00.04 0.31 0.29 0.03 0.02 0.12 0.000.58 4.55 4.24 0.42 0.27 1.76 0.004.31 4.92 3.65 9.86 19.15 6.54 0.00

URBAN,ADT>6K

29 156 179 23 15 111 10.06 0.32 0.37 0.05 0.03 0.23 0.000.34 1.81 2.07 0.27 0.17 1.29 0.016.58 5.12 4.66 16.20 31.91 12.51 14.29

TOTAL 441 3047 3839 142 47 887 70.92 6.33 7.98 0.30 0.10 1.84 0.01


J-9


MedianBarrier

End ofBridge

Side ofBridge

Pier atUnderpass

Top of Underpass

BridgeCrossing

Gate

AttenuationDevice

RURAL,ADT<=2K

7 62 163 5 0 0 00.01 0.13 0.34 0.01 0.00 0.00 0.000.05 0.42 1.11 0.03 0.00 0.00 0.00

14.29 40.26 33.00 22.73 0.00 0.00RURAL,ADT=2-6K

5 64 209 4 0 0 10.01 0.13 0.43 0.01 0.00 0.00 0.000.04 0.45 1.47 0.03 0.00 0.00 0.01

10.20 41.56 42.31 18.18 0.00 12.50RURAL,ADT>6K

7 17 55 3 0 0 40.01 0.04 0.11 0.01 0.00 0.00 0.010.10 0.25 0.81 0.04 0.00 0.00 0.06

14.29 11.04 11.13 13.64 0.00 50.00URBAN,ADT<=2K

1 1 7 0 0 0 00.00 0.00 0.01 0.00 0.00 0.00 0.000.17 0.17 1.16 0.00 0.00 0.00 0.002.04 0.65 1.42 0.00 0.00 0.00

URBAN,ADT=2-6K

2 5 19 5 0 0 00.00 0.01 0.04 0.01 0.00 0.00 0.000.06 0.15 0.58 0.15 0.00 0.00 0.004.08 3.25 3.85 22.73 0.00 0.00

URBAN,ADT>6K

27 5 41 5 1 0 30.06 0.01 0.09 0.01 0.00 0.00 0.010.31 0.06 0.47 0.06 0.01 0.00 0.03

55.10 3.25 8.30 22.73 100.00 37.50TOTAL 49 154 494 22 1 0 8

0.10 0.32 1.03 0.05 0.00 0.00 0.02


J-10


Rocks fromTrucks

Debris onRoad

Object from

AnotherVehicle

Previously WreckedVehicle

Other Machinery

Other Object

ConcreteTrafficBarrier

RURAL,ADT<=2K

0 12 6 19 5 37 30.00 0.02 0.01 0.04 0.01 0.08 0.010.00 0.08 0.04 0.13 0.03 0.25 0.02

52.17 37.50 24.05 27.78 37.76 15.00RURAL,ADT=2-6K

0 8 5 30 5 28 30.00 0.02 0.01 0.06 0.01 0.06 0.010.00 0.06 0.04 0.21 0.04 0.20 0.02

34.78 31.25 37.97 27.78 28.57 15.00RURAL,ADT>6K

0 2 3 12 1 15 40.00 0.00 0.01 0.02 0.00 0.03 0.010.00 0.03 0.04 0.18 0.01 0.22 0.06

8.70 18.75 15.19 5.56 15.31 20.00URBAN,ADT<=2K

0 0 0 1 0 0 00.00 0.00 0.00 0.00 0.00 0.00 0.000.00 0.00 0.00 0.17 0.00 0.00 0.00

0.00 0.00 1.27 0.00 0.00 0.00URBAN,ADT=2-6K

0 1 0 3 0 7 10.00 0.00 0.00 0.01 0.00 0.01 0.000.00 0.03 0.00 0.09 0.00 0.21 0.03

4.35 0.00 3.80 0.00 7.14 5.00URBAN,ADT>6K

0 0 2 14 7 11 90.00 0.00 0.00 0.03 0.01 0.02 0.020.00 0.00 0.02 0.16 0.08 0.13 0.10

0.00 12.50 17.72 38.89 11.22 45.00TOTAL 0 23 16 79 18 98 20

0.00 0.05 0.03 0.16 0.04 0.20 0.04


J-11


Delineator Post

RetainingWall

HOVLane Gate

GuardPost

FireHydrant

Ditch(Earth)

Embank-ment

TOTAL

RURAL,ADT<=2K

189 7 0 2 3 359 365 14,6220.39 0.01 0.00 0.00 0.01 0.75 0.76 30.381.29 0.05 0.00 0.01 0.02 2.46 2.50

60.19 36.84 33.33 9.68 40.56 49.06 RURAL,ADT=2-6K

80 3 0 1 5 265 246 14,1990.17 0.01 0.00 0.00 0.01 0.55 0.51 29.500.56 0.02 0.00 0.01 0.04 1.87 1.73

25.48 15.79 16.67 16.13 29.94 33.06 RURAL,ADT>6K

20 2 0 2 5 102 66 67680.04 0.00 0.00 0.00 0.01 0.21 0.14 14.060.30 0.03 0.00 0.03 0.07 1.51 0.98 6.37 10.53 33.33 16.13 11.53 8.87

URBAN,ADT<=2K

5 0 0 0 1 8 9 6020.01 0.00 0.00 0.00 0.00 0.02 0.02 1.250.83 0.00 0.00 0.00 0.17 1.33 1.50 1.59 0.00 0.00 3.23 0.90 1.21

URBAN,ADT=2-6K

4 1 0 1 7 60 27 33000.01 0.00 0.00 0.00 0.01 0.12 0.06 6.860.12 0.03 0.00 0.03 0.21 1.82 0.82 1.27 5.26 16.67 22.58 6.78 3.63

URBAN,ADT>6K

16 6 0 0 10 91 31 86380.03 0.01 0.00 0.00 0.02 0.19 0.06 17.950.19 0.07 0.00 0.00 0.12 1.05 0.36 5.10 31.58 0.00 32.26 10.28 4.17

TOTAL 314 19 0 6 31 885 744 48,129

0.65 0.04 0.00 0.01 0.06 1.84 1.55 100.00


J-12

Table J-5. Crashes by Alignment. Dev, ADT ALIGNMENTFrequencyPercentRow PercentCol Percent

UnknownStraight,

LevelStraight,

GradeStraight,Hillcrest

Curve,Level

Curve,Grade

Curve,Hillcrest

TOTAL

RURAL,ADT<=2K

0 9615 50 167 4726 22 42 14,6220.00 19.98 0.10 0.35 9.82 0.05 0.09 30.380.00 65.76 0.34 1.14 32.32 0.15 0.29

24.83 40.32 41.85 53.96 48.89 51.22 RURAL,ADT=2-6K

0 11,429 38 146 2546 15 25 14,1990.00 23.75 0.08 0.30 5.29 0.03 0.05 29.500.00 80.49 0.27 1.03 17.93 0.11 0.18

29.52 30.65 36.59 29.07 33.33 30.49 RURAL,ADT>6K

0 6090 17 45 607 4 5 67680.00 12.65 0.04 0.09 1.26 0.01 0.01 14.060.00 89.98 0.25 0.66 8.97 0.06 0.07

15.73 13.71 11.28 6.93 8.89 6.10 URBAN,ADT<=2K

0 504 1 3 91 1 2 6020.00 1.05 0.00 0.01 0.19 0.00 0.00 1.250.00 83.72 0.17 0.50 15.12 0.17 0.33

1.30 0.81 0.75 1.04 2.22 2.44 URBAN,ADT=2-6K

0 2908 4 15 367 0 6 33000.00 6.04 0.01 0.03 0.76 0.00 0.01 6.860.00 88.12 0.12 0.45 11.12 0.00 0.18

7.51 3.23 3.76 4.19 0.00 7.32 URBAN,ADT>6K

0 8174 14 23 422 3 2 86380.00 16.98 0.03 0.05 0.88 0.01 0.00 17.950.00 94.63 0.16 0.27 4.89 0.03 0.02

21.11 11.29 5.76 4.82 6.67 2.44 TOTAL 0 38,720 124 399 8759 45 82 48,129

0.00 80.45 0.26 0.83 18.20 0.09 0.17 100.00


J-13

Table J-6. Crashes by Degree of Curve. Dev, ADT DEGREE OF CURVEFrequencyPercentRow PercentCol Percent

Unknown No Curve 0.1 to 1.9 2.0 to 3.9 4.0 to 5.9 6.0 to 7.9 8.0 to 9.9

RURAL,ADT<=2K

601 8853 695 1384 1195 566 2841.25 18.39 1.44 2.88 2.48 1.18 0.594.11 60.55 4.75 9.47 8.17 3.87 1.94

38.87 25.54 26.28 37.89 50.66 56.54 59.29RURAL,ADT=2-6K

451 10,268 919 1153 640 233 1120.94 21.33 1.91 2.40 1.33 0.48 0.233.18 72.31 6.47 8.12 4.51 1.64 0.79

29.17 29.62 34.74 31.56 27.13 23.28 23.38RURAL,ADT>6K

147 5429 422 401 193 74 170.31 11.28 0.88 0.83 0.40 0.15 0.042.17 80.22 6.24 5.92 2.85 1.09 0.259.51 15.66 15.95 10.98 8.18 7.39 3.55

URBAN,ADT<=2K

18 451 21 44 35 9 50.04 0.94 0.04 0.09 0.07 0.02 0.012.99 74.92 3.49 7.31 5.81 1.50 0.831.16 1.30 0.79 1.20 1.48 0.90 1.04

URBAN,ADT=2-6K

119 2523 167 194 120 52 360.25 5.24 0.35 0.40 0.25 0.11 0.073.61 76.45 5.06 5.88 3.64 1.58 1.097.70 7.28 6.31 5.31 5.09 5.19 7.52

URBAN,ADT>6K

210 7143 421 477 176 67 250.44 14.84 0.87 0.99 0.37 0.14 0.052.43 82.69 4.87 5.52 2.04 0.78 0.29

13.58 20.60 15.92 13.06 7.46 6.69 5.22TOTAL 1546 34,667 2645 3653 2359 1001 479

3.21 72.03 5.50 7.59 4.90 2.08 1.00


J-14

Table J-6. Crashes by Degree of Curve (continued). Dev, ADT DEGREE OF CURVEFrequencyPercentRow PercentCol Percent

10.0 to11.9

12.0 to13.9

14.0 to15.9

16.0 to17.9

18.0 andOver

TOTAL 0 to 3.9 4 to 18+

RURAL,ADT<=2K

456 107 110 31 340 14,622 2079 30890.95 0.22 0.23 0.06 0.71 30.383.12 0.73 0.75 0.21 2.33 14.22 21.13

57.36 74.31 54.19 60.78 58.02RURAL,ADT=2-6K

210 20 52 12 129 14,199 2072 14080.44 0.04 0.11 0.02 0.27 29.501.48 0.14 0.37 0.08 0.91 14.59 9.92

26.42 13.89 25.62 23.53 22.01RURAL,ADT>6K

36 6 12 1 30 6768 823 3690.07 0.01 0.02 0.00 0.06 14.060.53 0.09 0.18 0.01 0.44 12.16 5.454.53 4.17 5.91 1.96 5.12

URBAN,ADT<=2K

14 1 1 0 3 602 65 680.03 0.00 0.00 0.00 0.01 1.252.33 0.17 0.17 0.00 0.50 10.80 11.301.76 0.69 0.49 0.00 0.51

URBAN,ADT=2-6K

32 6 9 4 38 3300 361 2970.07 0.01 0.02 0.01 0.08 6.860.97 0.18 0.27 0.12 1.15 10.94 9.004.03 4.17 4.43 7.84 6.48

URBAN,ADT>6K

47 4 19 3 46 8638 898 3870.10 0.01 0.04 0.01 0.10 17.950.54 0.05 0.22 0.03 0.53 10.40 4.485.91 2.78 9.36 5.88 7.85

TOTAL 795 144 203 51 586 48,129 6298 56181.65 0.30 0.42 0.11 1.22 100.00


J-15

Table J-7. Crashes by Weather.Dev, ADT WEATHERFrequencyPercentRow PercentCol Percent

Clear(Cloudy)

Raining Snowing Fog

RURAL, ADT<=2K 13,054 1057 57 39527.12 2.20 0.12 0.8289.28 7.23 0.39 2.7030.54 25.74 44.88 39.50

RURAL, ADT=2-6K 12,404 1347 44 35125.77 2.80 0.09 0.7387.36 9.49 0.31 2.4729.02 32.81 34.65 35.10

RURAL, ADT>6K 5997 637 8 10912.46 1.32 0.02 0.2388.61 9.41 0.12 1.6114.03 15.51 6.30 10.90

URBAN, ADT<=2K 541 47 0 141.12 0.10 0.00 0.03

89.87 7.81 0.00 2.331.27 1.14 0.00 1.40

URBAN, ADT=2-6K 2960 273 7 526.15 0.57 0.01 0.11

89.70 8.27 0.21 1.586.93 6.65 5.51 5.20

URBAN, ADT>6K 7787 745 11 7916.18 1.55 0.02 0.1690.15 8.62 0.13 0.9118.22 18.14 8.66 7.90

TOTAL 42,743 4106 127 100088.81 8.53 0.26 2.08


J-16

Table J-7. Crashes by Weather (continued). Dev, ADT WEATHERFrequencyPercentRow PercentCol Percent

BlowingDust

Smoke Other Sleeting TOTAL

RURAL, ADT<=2K 5 5 11 38 14,6220.01 0.01 0.02 0.08 30.380.03 0.03 0.08 0.26

55.56 38.46 50.00 34.86 RURAL, ADT=2-6K 0 5 6 42 14,199

0.00 0.01 0.01 0.09 29.500.00 0.04 0.04 0.30 0.00 38.46 27.27 38.53

RURAL, ADT>6K 2 2 2 11 67680.00 0.00 0.00 0.02 14.060.03 0.03 0.03 0.16

22.22 15.38 9.09 10.09URBAN, ADT<=2K 0 0 0 0 602

0.00 0.00 0.00 0.00 1.250.00 0.00 0.00 0.000.00 0.00 0.00 0.00

URBAN, ADT=2-6K 1 1 0 6 33000.00 0.00 0.00 0.01 6.860.03 0.03 0.00 0.18

11.11 7.69 0.00 5.50 URBAN, ADT>6K 1 0 3 12 8638

0.00 0.00 0.01 0.02 17.950.01 0.00 0.03 0.14

11.11 0.00 13.64 11.01 TOTAL 9 13 22 109 48,129

0.02 0.03 0.05 0.23 100.00


J-17

Table J-8. Crashes by Surface Condition. Dev, ADT SURFACE CONDITIONFrequencyPercentRow PercentCol Percent

Dry Wet Muddy Snowy Icy TOTAL

RURAL, ADT<=2K 12,539 1909 7 167 0 14,62226.05 3.97 0.01 0.35 0.00 30.3885.75 13.06 0.05 1.14 0.00 30.61 28.48 38.89 37.53

RURAL, ADT=2-6K 11,886 2129 2 182 0 14,19924.70 4.42 0.00 0.38 0.00 29.5083.71 14.99 0.01 1.28 0.00 29.02 31.76 11.11 40.90

RURAL, ADT>6K 5775 949 2 42 0 676812.00 1.97 0.00 0.09 0.00 14.0685.33 14.02 0.03 0.62 0.00 14.10 14.16 11.11 9.44

URBAN, ADT<=2K 513 88 0 1 0 6021.07 0.18 0.00 0.00 0.00 1.25

85.22 14.62 0.00 0.17 0.00 1.25 1.31 0.00 0.22

URBAN, ADT=2-6K 2846 437 1 16 0 33005.91 0.91 0.00 0.03 0.00 6.86

86.24 13.24 0.03 0.48 0.00 6.95 6.52 5.56 3.60

URBAN, ADT>6K 7403 1192 6 37 0 863815.38 2.48 0.01 0.08 0.00 17.9585.70 13.80 0.07 0.43 0.00 18.07 17.78 33.33 8.31

TOTAL 40,962 6704 18 445 0 48,12985.11 13.93 0.04 0.92 0.00 100.00


J-18

Table J-9. Crashes by Light Condition. Dev, ADT LIGHT CONDITIONFrequencyPercentRow PercentCol Percent

Daylight Dawn Dark NotLighted

DarkLighted

Dusk TOTAL

RURAL, ADT<=2K 8338 269 5344 389 282 14,62217.32 0.56 11.10 0.81 0.59 30.3857.02 1.84 36.55 2.66 1.93 28.03 31.54 39.65 12.36 30.99

RURAL, ADT=2-6K 8738 277 4336 609 239 14,19918.16 0.58 9.01 1.27 0.50 29.5061.54 1.95 30.54 4.29 1.68 29.38 32.47 32.17 19.35 26.26

RURAL, ADT>6K 4260 120 1776 464 148 67688.85 0.25 3.69 0.96 0.31 14.06

62.94 1.77 26.24 6.86 2.19 14.32 14.07 13.18 14.74 16.26

URBAN, ADT<=2K 377 7 164 49 5 6020.78 0.01 0.34 0.10 0.01 1.25

62.62 1.16 27.24 8.14 0.83 1.27 0.82 1.22 1.56 0.55

URBAN, ADT=2-6K 2146 51 647 386 70 33004.46 0.11 1.34 0.80 0.15 6.86

65.03 1.55 19.61 11.70 2.12 7.22 5.98 4.80 12.27 7.69

URBAN, ADT>6K 5883 129 1210 1250 166 863812.22 0.27 2.51 2.60 0.34 17.9568.11 1.49 14.01 14.47 1.92 19.78 15.12 8.98 39.72 18.24

TOTAL 29,742 853 13,477 3147 910 48,12961.80 1.77 28.00 6.54 1.89 100.00


J-19

Table J-10. Crashes by Month. Dev, ADT MONTHFrequencyPercentRow PercentCol Percent

January February March April May June July

RURAL, ADT<=2K 1058 984 1163 1197 1333 1192 13772.20 2.04 2.42 2.49 2.77 2.48 2.867.24 6.73 7.95 8.19 9.12 8.15 9.42

30.43 29.13 29.53 30.20 30.10 29.75 33.11RURAL, ADT=2-6K 1006 979 1157 1143 1308 1247 1207

2.09 2.03 2.40 2.37 2.72 2.59 2.517.09 6.89 8.15 8.05 9.21 8.78 8.50

28.93 28.98 29.38 28.84 29.53 31.12 29.02RURAL, ADT>6K 489 487 514 555 646 542 571

1.02 1.01 1.07 1.15 1.34 1.13 1.197.23 7.20 7.59 8.20 9.54 8.01 8.44

14.06 14.42 13.05 14.00 14.59 13.53 13.73URBAN, ADT<=2K 44 40 52 60 48 46 52

0.09 0.08 0.11 0.12 0.10 0.10 0.117.31 6.64 8.64 9.97 7.97 7.64 8.641.27 1.18 1.32 1.51 1.08 1.15 1.25

URBAN, ADT=2-6K 258 222 306 286 292 275 2620.54 0.46 0.64 0.59 0.61 0.57 0.547.82 6.73 9.27 8.67 8.85 8.33 7.947.42 6.57 7.77 7.22 6.59 6.86 6.30

URBAN, ADT>6K 622 666 746 722 802 705 6901.29 1.38 1.55 1.50 1.67 1.46 1.437.20 7.71 8.64 8.36 9.28 8.16 7.99

17.89 19.72 18.94 18.22 18.11 17.59 16.59TOTAL 3477 3378 3938 3963 4429 4007 4159

7.22 7.02 8.18 8.23 9.20 8.33 8.64


J-20

Table J-10. Crashes by Month (continued). Dev, ADT MONTHFrequencyPercentRow PercentCol Percent

August September October November December TOTAL

RURAL, ADT<=2K 1219 1198 1331 1287 1283 14,6222.53 2.49 2.77 2.67 2.67 30.388.34 8.19 9.10 8.80 8.77

30.04 30.38 30.30 31.05 30.27RURAL, ADT=2-6K 1180 1155 1283 1225 1309 14,199

2.45 2.40 2.67 2.55 2.72 29.508.31 8.13 9.04 8.63 9.22

29.08 29.29 29.21 29.55 30.88RURAL, ADT>6K 600 606 620 580 558 6768

1.25 1.26 1.29 1.21 1.16 14.068.87 8.95 9.16 8.57 8.24

14.79 15.37 14.11 13.99 13.16URBAN, ADT<=2K 34 52 63 56 55 602

0.07 0.11 0.13 0.12 0.11 1.255.65 8.64 10.47 9.30 9.140.84 1.32 1.43 1.35 1.30

URBAN, ADT=2-6K 283 269 298 280 269 33000.59 0.56 0.62 0.58 0.56 6.868.58 8.15 9.03 8.48 8.156.97 6.82 6.78 6.76 6.35

URBAN, ADT>6K 742 663 798 717 765 86381.54 1.38 1.66 1.49 1.59 17.958.59 7.68 9.24 8.30 8.86

18.28 16.81 18.17 17.30 18.05TOTAL 4058 3943 4393 4145 4239 48,129

8.43 8.19 9.13 8.61 8.81 100.00


J-21

Table J-11. Crashes by Day of Week. Dev, ADT DAY OF WEEKFrequencyPercentRow PercentCol Percent

Sunday Monday Tuesday Wednesday Thursday Friday Saturday TOTAL

RURAL, ADT<=2K 2449 1812 1653 1789 1848 2320 2751 14,6225.09 3.76 3.43 3.72 3.84 4.82 5.72 30.38

16.75 12.39 11.30 12.23 12.64 15.87 18.81 35.55 28.89 28.02 28.83 28.31 28.96 33.04

RURAL, ADT=2-6K

2111 1866 1663 1805 1920 2339 2495 14,1994.39 3.88 3.46 3.75 3.99 4.86 5.18 29.50

14.87 13.14 11.71 12.71 13.52 16.47 17.57 30.65 29.76 28.19 29.08 29.42 29.20 29.96

RURAL, ADT>6K 893 849 860 914 912 1162 1178 67681.86 1.76 1.79 1.90 1.89 2.41 2.45 14.06

13.19 12.54 12.71 13.50 13.48 17.17 17.41 12.96 13.54 14.58 14.73 13.97 14.51 14.15

URBAN, ADT<=2K 80 80 85 84 86 99 88 6020.17 0.17 0.18 0.17 0.18 0.21 0.18 1.25

13.29 13.29 14.12 13.95 14.29 16.45 14.62 1.16 1.28 1.44 1.35 1.32 1.24 1.06

URBAN, ADT=2-6K

392 471 425 447 510 571 484 33000.81 0.98 0.88 0.93 1.06 1.19 1.01 6.86

11.88 14.27 12.88 13.55 15.45 17.30 14.67 5.69 7.51 7.20 7.20 7.81 7.13 5.81

URBAN, ADT>6K 963 1193 1214 1167 1251 1519 1331 86382.00 2.48 2.52 2.42 2.60 3.16 2.77 17.95

11.15 13.81 14.05 13.51 14.48 17.59 15.41 13.98 19.02 20.58 18.80 19.17 18.96 15.98

TOTAL 6888 6271 5900 6206 6527 8010 8327 48,12914.31 13.03 12.26 12.89 13.56 16.64 17.30 100.00


J-22

Table J-12. Crashes by Time. Dev, ADT TIMEFrequencyPercentRow PercentCol Percent

MIDNIGHT-12:59 AM

1-1:59 AM 2-2:59 AM 3-3:59 AM 4-4:59 AM 5-5:59 AM

RURAL, ADT<=2K 512 474 415 288 272 3341.06 0.98 0.86 0.60 0.57 0.693.50 3.24 2.84 1.97 1.86 2.28

35.83 37.92 35.47 33.57 38.15 34.90RURAL, ADT=2-6K 440 391 343 268 257 320

0.91 0.81 0.71 0.56 0.53 0.663.10 2.75 2.42 1.89 1.81 2.25

30.79 31.28 29.32 31.24 36.04 33.44RURAL, ADT>6K 180 150 159 146 86 152

0.37 0.31 0.33 0.30 0.18 0.322.66 2.22 2.35 2.16 1.27 2.25

12.60 12.00 13.59 17.02 12.06 15.88URBAN, ADT<=2K 22 15 22 11 9 11

0.05 0.03 0.05 0.02 0.02 0.023.65 2.49 3.65 1.83 1.50 1.831.54 1.20 1.88 1.28 1.26 1.15

URBAN, ADT=2-6K 99 58 59 47 30 490.21 0.12 0.12 0.10 0.06 0.103.00 1.76 1.79 1.42 0.91 1.486.93 4.64 5.04 5.48 4.21 5.12

URBAN, ADT>6K 176 162 172 98 59 910.37 0.34 0.36 0.20 0.12 0.192.04 1.88 1.99 1.13 0.68 1.05

12.32 12.96 14.70 11.42 8.27 9.51TOTAL 1429 1250 1170 858 713 957

2.97 2.60 2.43 1.78 1.48 1.99


J-23

Table J-12. Crashes by Time (continued). Dev, ADT TIMEFrequencyPercentRow PercentCol Percent

6-6:59 AM 7-7:59 AM 8-8:59 AM 9-9:59 AM 10-10:59 AM 11-11:59 AM

RURAL, ADT<=2K 469 711 521 447 510 6050.97 1.48 1.08 0.93 1.06 1.263.21 4.86 3.56 3.06 3.49 4.14

31.23 29.13 28.13 29.25 29.29 29.89RURAL, ADT=2-6K 480 746 562 455 521 576

1.00 1.55 1.17 0.95 1.08 1.203.38 5.25 3.96 3.20 3.67 4.06

31.96 30.56 30.35 29.78 29.93 28.46RURAL, ADT>6K 207 333 259 251 255 263

0.43 0.69 0.54 0.52 0.53 0.553.06 4.92 3.83 3.71 3.77 3.89

13.78 13.64 13.98 16.43 14.65 12.99URBAN, ADT<=2K 25 32 30 18 26 23

0.05 0.07 0.06 0.04 0.05 0.054.15 5.32 4.98 2.99 4.32 3.821.66 1.31 1.62 1.18 1.49 1.14

URBAN, ADT=2-6K 87 171 118 96 118 1570.18 0.36 0.25 0.20 0.25 0.332.64 5.18 3.58 2.91 3.58 4.765.79 7.01 6.37 6.28 6.78 7.76

URBAN, ADT>6K 234 448 362 261 311 4000.49 0.93 0.75 0.54 0.65 0.832.71 5.19 4.19 3.02 3.60 4.63

15.58 18.35 19.55 17.08 17.86 19.76TOTAL 1502 2441 1852 1528 1741 2024

3.12 5.07 3.85 3.17 3.62 4.21


J-24


NOON-12:59PM

1-1:59 PM 2-2:59 PM 3-3:59 PM 4-4:59 PM 5-5:59 PM

RURAL, ADT<=2K 604 646 683 904 901 9381.25 1.34 1.42 1.88 1.87 1.954.13 4.42 4.67 6.18 6.16 6.41

25.99 27.13 26.09 28.50 26.59 26.91RURAL, ADT=2-6K 692 714 836 904 945 991

1.44 1.48 1.74 1.88 1.96 2.064.87 5.03 5.89 6.37 6.66 6.98

29.78 29.99 31.93 28.50 27.88 28.43RURAL, ADT>6K 308 370 355 464 495 515

0.64 0.77 0.74 0.96 1.03 1.074.55 5.47 5.25 6.86 7.31 7.61

13.25 15.54 13.56 14.63 14.61 14.77URBAN, ADT<=2K 28 28 32 38 38 45

0.06 0.06 0.07 0.08 0.08 0.094.65 4.65 5.32 6.31 6.31 7.481.20 1.18 1.22 1.20 1.12 1.29

URBAN, ADT=2-6K 185 172 210 211 249 2660.38 0.36 0.44 0.44 0.52 0.555.61 5.21 6.36 6.39 7.55 8.067.96 7.22 8.02 6.65 7.35 7.63

URBAN, ADT>6K 507 451 502 651 761 7311.05 0.94 1.04 1.35 1.58 1.525.87 5.22 5.81 7.54 8.81 8.46

21.82 18.94 19.17 20.52 22.46 20.97TOTAL 2324 2381 2618 3172 3389 3486

4.83 4.95 5.44 6.59 7.04 7.24


J-25


6-6:59 PM

7-7:59 PM

8-8:59 PM

9-9:59 PM

10-10:59PM

11-11:59PM

TOTAL

RURAL, ADT<=2K 934 743 727 742 674 568 14,6221.94 1.54 1.51 1.54 1.40 1.18 30.386.39 5.08 4.97 5.07 4.61 3.88

28.89 30.89 33.60 34.35 37.91 36.55RURAL, ADT=2-6K

918 660 607 617 518 438 14,1991.91 1.37 1.26 1.28 1.08 0.91 29.56.47 4.65 4.27 4.35 3.65 3.08

28.39 27.44 28.05 28.56 29.13 28.19RURAL, ADT>6K 486 344 319 281 194 196 6768

1.01 0.71 0.66 0.58 0.40 0.41 14.067.18 5.08 4.71 4.15 2.87 2.90

15.03 14.30 14.74 13.01 10.91 12.61URBAN, ADT<=2K 29 31 24 31 13 21 602

0.06 0.06 0.05 0.06 0.03 0.04 1.254.82 5.15 3.99 5.15 2.16 3.490.90 1.29 1.11 1.44 0.73 1.35

URBAN, ADT=2-6K

222 186 130 165 108 107 33000.46 0.39 0.27 0.34 0.22 0.22 6.866.73 5.64 3.94 5.00 3.27 3.246.87 7.73 6.01 7.64 6.07 6.89

URBAN, ADT>6K 644 441 357 324 271 224 86381.34 0.92 0.74 0.67 0.56 0.47 17.957.46 5.11 4.13 3.75 3.14 2.59

19.92 18.34 16.50 15.00 15.24 14.41TOTAL 3233 2405 2164 2160 1778 1554 48,120

6.72 5.00 4.50 4.49 3.69 3.23 100.00


J-26

Table J-13. Crashes by Vehicle Movement. Dev, ADT VEHICLE MOVEMENTS / MANNER OF COLLISIONFrequencyPercentRow PercentCol Percent

Single MotorVehicle Straight

Single MotorVehicle

Right Turn

SingleMotor

Vehicle Left Turn

Single MotorVehicle Backing

Single MotorVehicleOther

Angle BothStraight

RURAL, ADT<=2K 9891 71 98 11 3 132220.55 0.15 0.20 0.02 0.01 2.7567.64 0.49 0.67 0.08 0.02 9.0445.52 25.09 28.08 39.29 25.00 21.85

RURAL, ADT=2-6K 6806 68 74 10 4 158014.14 0.14 0.15 0.02 0.01 3.2847.93 0.48 0.52 0.07 0.03 11.1331.32 24.03 21.20 35.71 33.33 26.11

RURAL, ADT>6K 2101 34 38 4 5 7404.37 0.07 0.08 0.01 0.01 1.54

31.04 0.50 0.56 0.06 0.07 10.939.67 12.01 10.89 14.29 41.67 12.23

URBAN, ADT<=2K 242 5 4 0 0 1370.50 0.01 0.01 0.00 0.00 0.28

40.20 0.83 0.66 0.00 0.00 22.761.11 1.77 1.15 0.00 0.00 2.26

URBAN, ADT=2-6K 1027 27 46 1 0 6592.13 0.06 0.10 0.00 0.00 1.37

31.12 0.82 1.39 0.03 0.00 19.974.73 9.54 13.18 3.57 0.00 10.89

URBAN, ADT>6K 1664 78 89 2 0 16133.46 0.16 0.18 0.00 0.00 3.35

19.26 0.90 1.03 0.02 0.00 18.677.66 27.56 25.50 7.14 0.00 26.66

TOTAL 21,731 283 349 28 12 605145.15 0.59 0.73 0.06 0.02 12.57


J-27

Table J-13. Crashes by Vehicle Movement (continued). Dev, ADT VEHICLE MOVEMENTS / MANNER OF COLLISIONFrequencyPercentRow PercentCol Percent

Angle 1Straight 2Backing

Angle 1Straight 2Stopped

Angle 1Straight 2

Right Turn

Angle 1Straight 2Left Turn

Angle BothRight Turn

Angle 1Right Turn

2 LeftTurn

RURAL, ADT<=2K 28 46 60 351 0 10.06 0.10 0.12 0.73 0.00 0.000.19 0.31 0.41 2.40 0.00 0.01

31.46 28.05 17.80 15.93 25.00RURAL, ADT=2-6K 31 48 96 543 0 1

0.06 0.10 0.20 1.13 0.00 0.000.22 0.34 0.68 3.82 0.00 0.01

34.83 29.27 28.49 24.65 25.00RURAL, ADT>6K 12 23 65 405 0 0

0.02 0.05 0.14 0.84 0.00 0.000.18 0.34 0.96 5.98 0.00 0.00

13.48 14.02 19.29 18.38 0.00URBAN, ADT<=2K 0 6 2 39 0 0

0.00 0.01 0.00 0.08 0.00 0.000.00 1.00 0.33 6.48 0.00 0.000.00 3.66 0.59 1.77 0.00

URBAN, ADT=2-6K 11 13 31 194 0 00.02 0.03 0.06 0.40 0.00 0.000.33 0.39 0.94 5.88 0.00 0.00

12.36 7.93 9.20 8.81 0.00URBAN, ADT>6K 7 28 83 671 0 2

0.01 0.06 0.17 1.39 0.00 0.000.08 0.32 0.96 7.77 0.00 0.027.87 17.07 24.63 30.46 50.00

TOTAL 89 164 337 2203 0 40.18 0.34 0.70 4.58 0.00 0.01


J-28

Table J-13. Crashes by Vehicle Movement. Dev, ADT VEHICLE MOVEMENTS / MANNER OF COLLISIONFrequencyPercentRow PercentCol Percent

Angle 1 RightTurn

2 Stopped

Angle BothLeft Turn

Angle 1 LeftTurn 2

Stopped

SameDirection

Both StraightRear End

SameDirection

Both StraightSideswipe

SameDirection 1Straight 2Stopped

RURAL, ADT<=2K 6 0 4 374 49 3370.01 0.00 0.01 0.78 0.10 0.700.04 0.00 0.03 2.56 0.34 2.30

16.22 0.00 16.67 16.18 16.17 8.39RURAL, ADT=2-6K 9 7 8 690 81 892

0.02 0.01 0.02 1.43 0.17 1.850.06 0.05 0.06 4.86 0.57 6.28

24.32 30.43 33.33 29.86 26.73 22.21RURAL, ADT>6K 5 5 4 524 63 870

0.01 0.01 0.01 1.09 0.13 1.810.07 0.07 0.06 7.74 0.93 12.85

13.51 21.74 16.67 22.67 20.79 21.66URBAN, ADT<=2K 1 1 0 19 5 36

0.00 0.00 0.00 0.04 0.01 0.070.17 0.17 0.00 3.16 0.83 5.982.70 4.35 0.00 0.82 1.65 0.90

URBAN, ADT=2-6K 5 1 2 147 13 3580.01 0.00 0.00 0.31 0.03 0.740.15 0.03 0.06 4.45 0.39 10.85

13.51 4.35 8.33 6.36 4.29 8.91URBAN, ADT>6K 11 9 6 557 92 1524

0.02 0.02 0.01 1.16 0.19 3.170.13 0.10 0.07 6.45 1.07 17.64

29.73 39.13 25.00 24.10 30.36 37.94TOTAL 37 23 24 2311 303 4017

0.08 0.05 0.05 4.80 0.63 8.35


J-29


SameDirection 1Straight 2

Right Turn

SameDirection 1Straight 2Left Turn

SameDirection

Both RightTurn

SameDirection 1Right Turn2 Left Turn

SameDirection 1Right Turn 2 Stopped

SameDirectionBoth Left

TurnRURAL, ADT<=2K 107 531 2 0 1 0

0.22 1.10 0.00 0.00 0.00 0.000.73 3.63 0.01 0.00 0.01 0.00

23.26 24.50 12.50 12.50 0.00RURAL, ADT=2-6K 152 810 4 0 0 2

0.32 1.68 0.01 0.00 0.00 0.001.07 5.70 0.03 0.00 0.00 0.01

33.04 37.38 25.00 0.00 14.29RURAL, ADT>6K 74 384 3 0 0 5

0.15 0.80 0.01 0.00 0.00 0.011.09 5.67 0.04 0.00 0.00 0.07

16.09 17.72 18.75 0.00 35.71URBAN, ADT<=2K 7 29 0 0 0 0

0.01 0.06 0.00 0.00 0.00 0.001.16 4.82 0.00 0.00 0.00 0.001.52 1.34 0.00 0.00 0.00

URBAN, ADT=2-6K 39 146 2 0 2 00.08 0.30 0.00 0.00 0.00 0.001.18 4.42 0.06 0.00 0.06 0.008.48 6.74 12.50 25.00 0.00

URBAN, ADT>6K 81 267 5 0 5 70.17 0.55 0.01 0.00 0.01 0.010.94 3.09 0.06 0.00 0.06 0.08

17.61 12.32 31.25 62.50 50.00TOTAL 460 2167 16 0 8 14

0.96 4.50 0.03 0.00 0.02 0.03


J-30


SameDirection 1Left Turn 2

Stopped

OppositeDirections

BothStraight

OppositeDirections 1Straight 2Backing

OppositeDirections 1Straight 2Stopped

OppositeDirections 1

Straight2 Right Turn

OppositeDirections 1Straight 2Left Turn

RURAL, ADT<=2K 0 746 3 20 4 5250.00 1.55 0.01 0.04 0.01 1.090.00 5.10 0.02 0.14 0.03 3.590.00 22.54 11.54 28.17 36.36 13.36

RURAL, ADT=2-6K 0 1287 8 20 4 9210.00 2.67 0.02 0.04 0.01 1.910.00 9.06 0.06 0.14 0.03 6.490.00 38.88 30.77 28.17 36.36 23.43

RURAL, ADT>6K 1 695 10 11 2 6530.00 1.44 0.02 0.02 0.00 1.360.01 10.27 0.15 0.16 0.03 9.65

100.00 21.00 38.46 15.49 18.18 16.61URBAN, ADT<=2K 0 23 0 1 0 45

0.00 0.05 0.00 0.00 0.00 0.090.00 3.82 0.00 0.17 0.00 7.480.00 0.69 0.00 1.41 0.00 1.14

URBAN, ADT=2-6K 0 148 2 0 1 4180.00 0.31 0.00 0.00 0.00 0.870.00 4.48 0.06 0.00 0.03 12.670.00 4.47 7.69 0.00 9.09 10.63

URBAN, ADT>6K 0 411 3 19 0 13690.00 0.85 0.01 0.04 0.00 2.840.00 4.76 0.03 0.22 0.00 15.850.00 12.42 11.54 26.76 0.00 34.83

TOTAL 1 3310 26 71 11 39310.00 6.88 0.05 0.15 0.02 8.17


J-31


OppositeDirections 1Backing 2Stopped

OppositeDirections 1Right Turn 2 Left Turn

OppositeDirections 1Right Turn 2 Stopped

OppositeDirections

Both Left Turn

OppositeDirections 1Left Turn2 Stopped

Other 1Straight2 Parked

RURAL, ADT<=2K 0 2 1 1 0 200.00 0.00 0.00 0.00 0.00 0.040.00 0.01 0.01 0.01 0.00 0.140.00 13.33 25.00 14.29 21.05

RURAL, ADT=2-6K 4 2 0 1 0 340.01 0.00 0.00 0.00 0.00 0.070.03 0.01 0.00 0.01 0.00 0.24

36.36 13.33 0.00 14.29 35.79RURAL, ADT>6K 2 3 0 1 0 22

0.00 0.01 0.00 0.00 0.00 0.050.03 0.04 0.00 0.01 0.00 0.33

18.18 20.00 0.00 14.29 23.16URBAN, ADT<=2K 0 0 0 0 0 0

0.00 0.00 0.00 0.00 0.00 0.000.00 0.00 0.00 0.00 0.00 0.000.00 0.00 0.00 0.00 0.00

URBAN, ADT=2-6K 0 0 1 0 0 50.00 0.00 0.00 0.00 0.00 0.010.00 0.00 0.03 0.00 0.00 0.150.00 0.00 25.00 0.00 5.26

URBAN, ADT>6K 5 8 2 4 0 140.01 0.02 0.00 0.01 0.00 0.030.06 0.09 0.02 0.05 0.00 0.16

45.45 53.33 50.00 57.14 14.74TOTAL 11 15 4 7 0 95

0.02 0.03 0.01 0.01 0.00 0.20


J-32


Other 1RightTurn

2 Parked

Other 1Left Turn2 Parked

Other 1Parked

2 Stopped

OtherBoth

Parked

OtherBoth

Backing

Other AllOthers

TOTAL

RURAL, ADT<=2K 0 0 0 0 0 7 14,6220.00 0.00 0.00 0.00 0.00 0.01 30.380.00 0.00 0.00 0.00 0.00 0.05

43.75 RURAL, ADT=2-6K 0 0 0 0 0 2 14,199

0.00 0.00 0.00 0.00 0.00 0.00 29.500.00 0.00 0.00 0.00 0.00 0.01

12.50 RURAL, ADT>6K 0 0 0 0 0 4 6768

0.00 0.00 0.00 0.00 0.00 0.01 14.060.00 0.00 0.00 0.00 0.00 0.06

25.00 URBAN, ADT<=2K 0 0 0 0 0 0 602

0.00 0.00 0.00 0.00 0.00 0.00 1.250.00 0.00 0.00 0.00 0.00 0.00

0.00 URBAN, ADT=2-6K 0 0 0 0 0 1 3300

0.00 0.00 0.00 0.00 0.00 0.00 6.860.00 0.00 0.00 0.00 0.00 0.03

6.25 URBAN, ADT>6K 0 0 0 0 0 2 8638

0.00 0.00 0.00 0.00 0.00 0.00 17.950.00 0.00 0.00 0.00 0.00 0.02

12.50 TOTAL 0 0 0 0 0 16 48,129

0.00 0.00 0.00 0.00 0.00 0.03 100.00


J-33

Table J-14. Crashes by Other Factors. Dev, ADT OTHER FACTORFrequencyPercentRow PercentCol Percent

No Code Applicable

Lost Control,Skidded

PassengerInterfered

AttentionDiverted

ObjectProjecting

fromVehicle

FootSlipped

off Brake

RURAL, ADT<=2K 10,331 139 43 597 4 721.47 0.29 0.09 1.24 0.01 0.0170.65 0.95 0.29 4.08 0.03 0.0534.48 36.29 50.59 41.34 40.00 38.89

RURAL, ADT=2-6K 8932 152 23 431 5 418.56 0.32 0.05 0.90 0.01 0.0162.91 1.07 0.16 3.04 0.04 0.0329.81 39.69 27.06 29.85 50.00 22.22

RURAL, ADT>6K 3632 38 5 172 0 17.55 0.08 0.01 0.36 0.00 0.00

53.66 0.56 0.07 2.54 0.00 0.0112.12 9.92 5.88 11.91 0.00 5.56

URBAN, ADT<=2K 407 1 2 12 0 00.85 0.00 0.00 0.02 0.00 0.00

67.61 0.17 0.33 1.99 0.00 0.001.36 0.26 2.35 0.83 0.00 0.00

URBAN, ADT=2-6K 2081 11 5 82 0 24.32 0.02 0.01 0.17 0.00 0.00

63.06 0.33 0.15 2.48 0.00 0.066.95 2.87 5.88 5.68 0.00 11.11

URBAN, ADT>6K 4581 42 7 150 1 49.52 0.09 0.01 0.31 0.00 0.01

53.03 0.49 0.08 1.74 0.01 0.0515.29 10.97 8.24 10.39 10.00 22.22

TOTAL 29,964 383 85 1444 10 1862.26 0.80 0.18 3.00 0.02 0.04


J-34

Table J-14. Crashes by Other Factors (continued). Dev, ADT OTHER FACTORFrequencyPercentRow PercentCol Percent

Gusty Winds

Vehicle Passing on Left

VehiclePassing on Right

VehicleChanging

Lane

ImproperlyParkedVehicle

VehicleForward

fromParking

RURAL, ADT<=2K 16 302 24 18 16 210.03 0.63 0.05 0.04 0.03 0.040.11 2.07 0.16 0.12 0.11 0.14

30.19 28.93 12.12 10.34 26.23 23.08RURAL, ADT=2-6K 22 436 64 33 23 31

0.05 0.91 0.13 0.07 0.05 0.060.15 3.07 0.45 0.23 0.16 0.22

41.51 41.76 32.32 18.97 37.70 34.07RURAL, ADT>6K 3 165 62 36 8 19

0.01 0.34 0.13 0.07 0.02 0.040.04 2.44 0.92 0.53 0.12 0.285.66 15.80 31.31 20.69 13.11 20.88

URBAN, ADT<=2K 0 13 4 2 2 10.00 0.03 0.01 0.00 0.00 0.000.00 2.16 0.66 0.33 0.33 0.170.00 1.25 2.02 1.15 3.28 1.10

URBAN, ADT=2-6K 3 57 11 10 5 50.01 0.12 0.02 0.02 0.01 0.010.09 1.73 0.33 0.30 0.15 0.155.66 5.46 5.56 5.75 8.20 5.49

URBAN, ADT>6K 9 71 33 75 7 140.02 0.15 0.07 0.16 0.01 0.030.10 0.82 0.38 0.87 0.08 0.16

16.98 6.80 16.67 43.10 11.48 15.38TOTAL 53 1044 198 174 61 91

0.11 2.17 0.41 0.36 0.13 0.1


J-35


VehicleBackward

fromParking

VehicleEntering

Driveway

VehicleLeaving

Driveway

VisionObstructed by

StandingVehicle

VisionObstructedby Moving

Vehicle

VisionObstructedby Embank-

mentRURAL, ADT<=2K 0 446 210 12 13 1

0.00 0.93 0.44 0.02 0.03 0.000.00 3.05 1.44 0.08 0.09 0.010.00 21.17 15.20 12.63 10.32 33.33

RURAL, ADT=2-6K 1 714 308 17 26 20.00 1.48 0.64 0.04 0.05 0.000.01 5.03 2.17 0.12 0.18 0.01

33.33 33.89 22.29 17.89 20.63 66.67RURAL, ADT>6K 0 377 274 14 30 0

0.00 0.78 0.57 0.03 0.06 0.000.00 5.57 4.05 0.21 0.44 0.000.00 17.89 19.83 14.74 23.81 0.00

URBAN, ADT<=2K 0 23 14 1 3 00.00 0.05 0.03 0.00 0.01 0.000.00 3.82 2.33 0.17 0.50 0.000.00 1.09 1.01 1.05 2.38 0.00

URBAN, ADT=2-6K 1 143 101 12 8 00.00 0.30 0.21 0.02 0.02 0.000.03 4.33 3.06 0.36 0.24 0.00

33.33 6.79 7.31 12.63 6.35 0.00URBAN, ADT>6K 1 404 475 39 46 0

0.00 0.84 0.99 0.08 0.10 0.000.01 4.68 5.50 0.45 0.53 0.00

33.33 19.17 34.37 41.05 36.51 0.00TOTAL 3 2107 1382 95 126 3

0.01 4.38 2.87 0.20 0.26 0.01


J-36


VisionObstructed byCommercial

Sign

VisionObstructedby Highway

Sign

VisionObstructed

by Glare

VisionObstructed by

Hillcrest

VisionObstructed

by Trees

VisionObstructedby Other

RURAL, ADT<=2K 1 2 81 15 5 1560.00 0.00 0.17 0.03 0.01 0.320.01 0.01 0.55 0.10 0.03 1.07

100.00 50.00 30.68 55.56 35.71 40.52RURAL, ADT=2-6K 0 1 81 8 5 108

0.00 0.00 0.17 0.02 0.01 0.220.00 0.01 0.57 0.06 0.04 0.760.00 25.00 30.68 29.63 35.71 28.05

RURAL, ADT>6K 0 1 41 1 1 520.00 0.00 0.09 0.00 0.00 0.110.00 0.01 0.61 0.01 0.01 0.770.00 25.00 15.53 3.70 7.14 13.51

URBAN, ADT<=2K 0 0 1 1 0 70.00 0.00 0.00 0.00 0.00 0.010.00 0.00 0.17 0.17 0.00 1.160.00 0.00 0.38 3.70 0.00 1.82

URBAN, ADT=2-6K 0 0 23 1 1 230.00 0.00 0.05 0.00 0.00 0.050.00 0.00 0.70 0.03 0.03 0.700.00 0.00 8.71 3.70 7.14 5.97

URBAN, ADT>6K 0 0 37 1 2 390.00 0.00 0.08 0.00 0.00 0.080.00 0.00 0.43 0.01 0.02 0.450.00 0.00 14.02 3.70 14.29 10.13

TOTAL 1 4 264 27 14 3850.00 0.01 0.55 0.06 0.03 0.80


J-37


Swerved to ChangeLanes

Swerved, Not Specified

Swerved,Surface orVisibility

Swerved,TrafficControl

Swerved, Pedestrianor Cyclist

Swerved,Animal

RURAL, ADT<=2K 25 172 2 3 5 5870.05 0.36 0.00 0.01 0.01 1.220.17 1.18 0.01 0.02 0.03 4.01

18.38 33.33 100.00 60.00 25.00 55.12RURAL, ADT=2-6K 50 195 0 1 10 298

0.10 0.41 0.00 0.00 0.02 0.620.35 1.37 0.00 0.01 0.07 2.10

36.76 37.79 0.00 20.00 50.00 27.98RURAL, ADT>6K 17 71 0 1 2 78

0.04 0.15 0.00 0.00 0.00 0.160.25 1.05 0.00 0.01 0.03 1.15

12.50 13.76 0.00 20.00 10.00 7.32URBAN, ADT<=2K 2 2 0 0 0 12

0.00 0.00 0.00 0.00 0.00 0.020.33 0.33 0.00 0.00 0.00 1.991.47 0.39 0.00 0.00 0.00 1.13

URBAN, ADT=2-6K 9 31 0 0 0 470.02 0.06 0.00 0.00 0.00 0.100.27 0.94 0.00 0.00 0.00 1.426.62 6.01 0.00 0.00 0.00 4.41

URBAN, ADT>6K 33 45 0 0 3 430.07 0.09 0.00 0.00 0.01 0.090.38 0.52 0.00 0.00 0.03 0.50

24.26 8.72 0.00 0.00 15.00 4.04TOTAL 136 516 2 5 20 1065

0.28 1.07 0.00 0.01 0.04 2.21


J-38


Swerved,Object in

Road

Swerved, SlowVehicle

Swerved,Vehicle Entering

Road

Swerved,AvoidingVehicle in

Wrong Lane

Swerved, AvoidingPreviousAccident

Slowed, NotSpecified

RURAL, ADT<=2K 37 58 81 181 3 990.08 0.12 0.17 0.38 0.01 0.210.25 0.40 0.55 1.24 0.02 0.68

40.22 18.01 31.15 41.61 75.00 21.57RURAL, ADT=2-6K 35 129 85 153 0 136

0.07 0.27 0.18 0.32 0.00 0.280.25 0.91 0.60 1.08 0.00 0.96

38.04 40.06 32.69 35.17 0.00 29.63RURAL, ADT>6K 14 67 34 42 1 76

0.03 0.14 0.07 0.09 0.00 0.160.21 0.99 0.50 0.62 0.01 1.12

15.22 20.81 13.08 9.66 25.00 16.56URBAN, ADT<=2K 0 4 5 4 0 4

0.00 0.01 0.01 0.01 0.00 0.010.00 0.66 0.83 0.66 0.00 0.660.00 1.24 1.92 0.92 0.00 0.87

URBAN, ADT=2-6K 3 20 21 17 0 260.01 0.04 0.04 0.04 0.00 0.050.09 0.61 0.64 0.52 0.00 0.793.26 6.21 8.08 3.91 0.00 5.66

URBAN, ADT>6K 3 44 34 38 0 1180.01 0.09 0.07 0.08 0.00 0.250.03 0.51 0.39 0.44 0.00 1.373.26 13.66 13.08 8.74 0.00 25.71

TOTAL 92 322 260 435 4 4590.19 0.67 0.54 0.90 0.01 0.95


J-39


Slowed,Surface orVisibility

Slowed, Traffic Control

Slowed, Pedestrianor Cyclist

Slowed, Animal

Slowed,Object in

Road

Slowed,Slow

Vehicle

RURAL, ADT<=2K 4 49 1 14 3 750.01 0.10 0.00 0.03 0.01 0.160.03 0.34 0.01 0.10 0.02 0.51

44.44 5.56 11.11 31.82 25.00 5.70RURAL, ADT=2-6K 4 118 2 18 5 241

0.01 0.25 0.00 0.04 0.01 0.500.03 0.83 0.01 0.13 0.04 1.70

44.44 13.38 22.22 40.91 41.67 18.33RURAL, ADT>6K 1 118 2 6 0 286

0.00 0.25 0.00 0.01 0.00 0.590.01 1.74 0.03 0.09 0.00 4.23

11.11 13.38 22.22 13.64 0.00 21.75URBAN, ADT<=2K 0 12 0 1 1 11

0.00 0.02 0.00 0.00 0.00 0.020.00 1.99 0.00 0.17 0.17 1.830.00 1.36 0.00 2.27 8.33 0.84

URBAN, ADT=2-6K 0 111 2 2 0 880.00 0.23 0.00 0.00 0.00 0.180.00 3.36 0.06 0.06 0.00 2.670.00 12.59 22.22 4.55 0.00 6.69

URBAN, ADT>6K 0 474 2 3 3 6140.00 0.98 0.00 0.01 0.01 1.280.00 5.49 0.02 0.03 0.03 7.110.00 53.74 22.22 6.82 25.00 46.69

TOTAL 9 882 9 44 12 13150.02 1.83 0.02 0.09 0.02 2.73


J-40


Slowed,Vehicle Entering

Road

Slowed,Vehicle in

Wrong Lane

Slowed,AvoidingPreviousAccident

Slowed,TurningRight

Slowed,Turning

Left

SchoolBus Related

Accident

RURAL, ADT<=2K 9 5 1 45 265 460.02 0.01 0.00 0.09 0.55 0.100.06 0.03 0.01 0.31 1.81 0.31

20.45 23.81 3.33 21.53 12.48 16.91RURAL, ADT=2-6K 15 7 11 65 657 97

0.03 0.01 0.02 0.14 1.37 0.200.11 0.05 0.08 0.46 4.63 0.68

34.09 33.33 36.67 31.10 30.95 35.66RURAL, ADT>6K 8 4 6 37 560 50

0.02 0.01 0.01 0.08 1.16 0.100.12 0.06 0.09 0.55 8.27 0.74

18.18 19.05 20.00 17.70 26.38 18.38URBAN, ADT<=2K 0 0 0 2 22 2

0.00 0.00 0.00 0.00 0.05 0.000.00 0.00 0.00 0.33 3.65 0.330.00 0.00 0.00 0.96 1.04 0.74

URBAN, ADT=2-6K 3 2 2 13 162 240.01 0.00 0.00 0.03 0.34 0.050.09 0.06 0.06 0.39 4.91 0.736.82 9.52 6.67 6.22 7.63 8.82

URBAN, ADT>6K 9 3 10 47 457 530.02 0.01 0.02 0.10 0.95 0.110.10 0.03 0.12 0.54 5.29 0.61

20.45 14.29 33.33 22.49 21.53 19.49TOTAL 44 21 30 209 2123 272

0.09 0.04 0.06 0.43 4.41 0.57


J-41


HighwayConstructionUnrelated toConstruction

HighwayConstruction

Related toConstruction

OtherConstruction

AreaUnrelated toConstruction

OtherConstructionArea Related

toConstruction

Accident onBeach

TOTAL

RURAL, ADT<=2K 302 85 0 5 0 14,6220.63 0.18 0.00 0.01 0.00 30.382.07 0.58 0.00 0.03 0.00

19.09 27.42 0.00 35.71 RURAL, ADT=2-6K

358 79 0 3 0 14,1990.74 0.16 0.00 0.01 0.00 29.502.52 0.56 0.00 0.02 0.00

22.63 25.48 0.00 21.43 RURAL, ADT>6K 292 60 0 3 0 6768

0.61 0.12 0.00 0.01 0.00 14.064.31 0.89 0.00 0.04 0.00

18.46 19.35 0.00 21.43 URBAN, ADT<=2K 23 1 0 0 0 602

0.05 0.00 0.00 0.00 0.00 1.253.82 0.17 0.00 0.00 0.00 1.45 0.32 0.00 0.00

URBAN, ADT=2-6K

111 19 0 2 0 33000.23 0.04 0.00 0.00 0.00 6.863.36 0.58 0.00 0.06 0.00 7.02 6.13 0.00 14.29

URBAN, ADT>6K 496 66 1 1 0 86381.03 0.14 0.00 0.00 0.00 17.955.74 0.76 0.01 0.01 0.00

31.35 21.29 100.00 7.14 TOTAL 1582 310 1 14 0 48,129

3.29 0.64 0.00 0.03 0.00 100.00

APPENDIX K: CRASHES BY DISTRICT GROUP

K-1

CRASHES BY THREE DISTRICT GROUPS

Tables K-1 to K-14 provide the crash frequencies and distributions for on-system, rural two-lanehighways by three district groups. The source of the data was the Traffic Accident Databaseprovided by the Texas Department of Public Safety and the Texas Department of Transportation. The data are from 1997 to 1999 for low-volume (ADT � 2000), rural two-lane highways.

Table K-1. Crashes by Accident Severity.District Accident Severity

FrequencyPercent

Row PercentCol Percent

Non-incapacitatingInjury

IncapacitatingInjury

Fatal TOTAL

HIGH-RATEDISTRICT

4107 1949 578 663428.09 13.33 3.95 45.3761.91 29.38 8.71 45.99 44.83 43.01

MID-RATE DISTRICT 3204 1581 476 526121.91 10.81 3.26 35.9860.90 30.05 9.0535.88 36.36 35.42

LOW-RATE DISTRICT 1619 818 290 272711.07 5.59 1.98 18.6559.37 30.00 10.6318.13 18.81 21.58

TOTAL 8930 4348 1344 1462261.07 29.74 9.19 100

APPENDIX K: CRASHES BY DISTRICT GROUPS

K-2

Table K-2. Crashes by Intersection Related.District Intersection Related

FrequencyPercent


IntersectionIntersection

RelatedDriveway

AccessNon

IntersectionTOTAL

HIGH-RATEDISTRICT

1025 578 696 4335 07.01 3.95 4.76 29.65 0.00

15.45 8.71 10.49 65.35 0.0045.60 42.94 54.85 44.42

MID-RATEDISTRICT

874 535 426 3426 05.98 3.66 2.91 23.43 0.00

16.61 10.17 8.10 65.12 0.0038.88 39.75 33.57 35.11

LOW-RATEDISTRICT

349 233 147 1998 02.39 1.59 1.01 13.66 0.00

12.80 8.54 5.39 73.27 0.0015.52 17.31 11.58 20.47

TOTAL 2248 1346 1269 9759 015.37 9.21 8.68 66.74 0.00

Table K-3. Crashes by First Harmful Event.District First Harmful Event

FrequencyPercent


Other Non-

CollisionOverturned Pedestrian

Other MotorVehicle

in TransitRR Train Parked Car

HIGH-RATEDISTRICT

44 1746 59 2139 9 390.30 11.94 0.40 14.63 0.06 0.270.66 26.32 0.89 32.24 0.14 0.59

36.07 41.48 40.69 47.03 25.71 54.17MID-RATEDISTRICT

43 1403 65 1677 16 190.29 9.60 0.44 11.47 0.11 0.130.82 26.67 1.24 31.88 0.30 0.36

35.25 33.33 44.83 36.87 45.71 26.39LOW-RATEDISTRICT

35 1060 21 732 10 140.24 7.25 0.14 5.01 0.07 0.101.28 38.87 0.77 26.84 0.37 0.51

28.69 25.18 14.48 16.09 28.57 19.44TOTAL 122 4209 145 4548 35 72

0.83 28.79 0.99 31.10 0.24 0.49


K-3

Table K-3. Crashes by First Harmful Event (continued).District First Harmful Event

FrequencyPercent


Pedalcyclist Animal Fixed Object Other Object TOTAL

HIGH-RATE DISTRICT 29 240 2301 28 00.20 1.64 15.74 0.19 0.000.44 3.62 34.68 0.42 0.00

40.85 37.68 48.55 63.64 MID-RATE DISTRICT 26 241 1760 11 0

0.18 1.65 12.04 0.08 0.000.49 4.58 33.45 0.21 0.00

36.62 37.83 37.14 25.00 LOW-RATE DISTRICT 16 156 678 5 0

0.11 1.07 4.64 0.03 0.000.59 5.72 24.86 0.18 0.00

22.54 24.49 14.31 11.36 TOTAL 71 637 4739 44 0

0.49 4.36 32.41 0.30 0.00

Table K-4. Crashes by Object Struck.District Object Struck

FrequencyPercent


No Code Applicable

VehicleOver-turned

Hole in Road

Vehicle Jack-knifed

Person Fell or

Jumpedfrom

Vehicle

Vehicle HitTrain onParallelTracks

TrainMoving

Forward

HIGH-RATEDISTRICT

3218 100 3 9 30 1 922.01 0.68 0.02 0.06 0.21 0.01 0.0648.51 1.51 0.05 0.14 0.45 0.02 0.1444.32 43.48 37.50 27.27 44.12 100.00 26.47

MID-RATEDISTRICT

2577 72 4 12 24 0 1517.62 0.49 0.03 0.08 0.16 0.00 0.1048.98 1.37 0.08 0.23 0.46 0.00 0.2935.49 31.30 50.00 36.36 35.29 0.00 44.12

LOW-RATEDISTRICT

1466 58 1 12 14 0 1010.03 0.40 0.01 0.08 0.10 0.00 0.0753.76 2.13 0.04 0.44 0.51 0.00 0.3720.19 25.22 12.50 36.36 20.59 0.00 29.41

TOTAL 7261 230 8 33 68 1 3449.66 1.57 0.05 0.23 0.47 0.01 0.23


K-4

Table K-4. Crashes by Object Struck (continued).District Object Struck

FrequencyPercent


TrainBacking

TrainStanding

Still

Train/Action

Unknown

Highway Sign

CurbCulvert/Headwall

Guardrail

HIGH-RATEDISTRICT

0 0 0 241 6 383 670.00 0.00 0.00 1.65 0.04 2.62 0.460.00 0.00 0.00 3.63 0.09 5.77 1.01

0.00 44.14 42.86 52.68 38.73MID-RATEDISTRICT

0 1 0 219 5 236 610.00 0.01 0.00 1.50 0.03 1.61 0.420.00 0.02 0.00 4.16 0.10 4.49 1.16

100.00 40.11 35.71 32.46 35.26LOW-RATEDISTRICT

0 0 0 86 3 108 450.00 0.00 0.00 0.59 0.02 0.74 0.310.00 0.00 0.00 3.15 0.11 3.96 1.65

0.00 15.75 21.43 14.86 26.01TOTAL 0 1 0 546 14 727 173

0.00 0.01 0.00 3.73 0.10 4.97 1.18


FrequencyPercent


RR SignalPole

RRCrossing

Gates

SignalPole/Post

SignalLight/Wires

Work ZoneBarricade

LuminairePole

UtilityPole

HIGH-RATE

DISTRICT

4 2 5 0 1 9 1120.03 0.01 0.03 0.00 0.01 0.06 0.770.06 0.03 0.08 0.00 0.02 0.14 1.69


3 1 7 0 1 4 1240.02 0.01 0.05 0.00 0.01 0.03 0.850.06 0.02 0.13 0.00 0.02 0.08 2.36


1 0 1 0 1 3 640.01 0.00 0.01 0.00 0.01 0.02 0.440.04 0.00 0.04 0.00 0.04 0.11 2.35

12.50 0.00 7.69 33.33 18.75 21.33TOTAL 8 3 13 0 3 16 300

0.05 0.02 0.09 0.00 0.02 0.11 2.05


K-5


FrequencyPercent


Mailbox Tree/Shrub FenceHouse/

BuildingCommercial

SignOther Fixed

Object

MaintenanceBarricade or

Materials

HIGH-RATE

DISTRICT

89 879 837 16 2 128 10.61 6.01 5.72 0.11 0.01 0.88 0.011.34 13.25 12.62 0.24 0.03 1.93 0.02

53.94 62.30 42.44 36.36 28.57 37.87 50.00MID-RATEDISTRICT

67 460 762 16 4 136 10.46 3.15 5.21 0.11 0.03 0.93 0.011.27 8.74 14.48 0.30 0.08 2.59 0.02

40.61 32.60 38.64 36.36 57.14 40.24 50.00LOW-RATEDISTRICT

9 72 373 12 1 74 00.06 0.49 2.55 0.08 0.01 0.51 0.000.33 2.64 13.68 0.44 0.04 2.71 0.005.45 5.10 18.91 27.27 14.29 21.89 0.00

TOTAL 165 1411 1972 44 7 338 21.13 9.65 13.49 0.30 0.05 2.31 0.01


FrequencyPercent


MedianBarrier

End ofBridge

Side ofBridge

Pier atUnderpass

Top of Underpass

BridgeCrossing

Gate

AttenuationDevice

HIGH-RATEDISTRICT

2 23 57 1 0 0 00.01 0.16 0.39 0.01 0.00 0.00 0.000.03 0.35 0.86 0.02 0.00 0.00 0.00

28.57 37.10 34.97 20.00MID-RATEDISTRICT

2 29 69 2 0 0 00.01 0.20 0.47 0.01 0.00 0.00 0.000.04 0.55 1.31 0.04 0.00 0.00 0.00

28.57 46.77 42.33 40.00LOW-RATEDISTRICT

3 10 37 2 0 0 00.02 0.07 0.25 0.01 0.00 0.00 0.000.11 0.37 1.36 0.07 0.00 0.00 0.00

42.86 16.13 22.70 40.00TOTAL 7 62 163 5 0 0 0

0.05 0.42 1.11 0.03 0.00 0.00 0.00


K-6


FrequencyPercent


Rocks fromTrucks

Debris onRoad

Object from AnotherVehicle

Previously WreckedVehicle

Other Machinery

Other Object

ConcreteTrafficBarrier

HIGH-RATE

DISTRICT

0 8 4 9 2 18 00.00 0.05 0.03 0.06 0.01 0.12 0.000.00 0.12 0.06 0.14 0.03 0.27 0.00


0 3 1 6 1 17 20.00 0.02 0.01 0.04 0.01 0.12 0.010.00 0.06 0.02 0.11 0.02 0.32 0.04


0 1 1 4 2 2 10.00 0.01 0.01 0.03 0.01 0.01 0.010.00 0.04 0.04 0.15 0.07 0.07 0.04

8.33 16.67 21.05 40.00 5.41 33.33TOTAL 0 12 6 19 5 37 3

0.00 0.08 0.04 0.13 0.03 0.25 0.02


FrequencyPercent


Delineator Post

RetainingWall

HOVLane Gate

GuardPost

FireHydrant

Ditch(Earth)

Embank-ment

TOTAL

HIGH-RATE

DISTRICT

46 3 0 0 0 145 164 66340.31 0.02 0.00 0.00 0.00 0.99 1.12 45.370.69 0.05 0.00 0.00 0.00 2.19 2.47

24.34 42.86 0.00 0.00 40.39 44.93 MID-RATEDISTRICT

61 2 0 0 3 128 123 52610.42 0.01 0.00 0.00 0.02 0.88 0.84 35.981.16 0.04 0.00 0.00 0.06 2.43 2.34

32.28 28.57 0.00 100.00 35.65 33.70 LOW-RATEDISTRICT

82 2 0 2 0 86 78 27270.56 0.01 0.00 0.01 0.00 0.59 0.53 18.653.01 0.07 0.00 0.07 0.00 3.15 2.86

43.39 28.57 100.00 0.00 23.96 21.37 TOTAL 189 7 0 2 3 359 365 14,622

1.29 0.05 0.00 0.01 0.02 2.46 2.50 100.00


K-7

Table K-5. Crashes by Alignment.District AlignmentPercent

FrequencyRow PercentCol Percent

Straight, Level

Straight,Grade

Straight,Hillcrest

Curve,Level

Curve,Grade

Curve,Hillcrest

TOTAL

HIGH-RATEDISTRICT

4021 23 97 2462 10 21 663427.50 0.16 0.66 16.84 0.07 0.14 45.3760.61 0.35 1.46 37.11 0.15 0.32 41.82 46.00 58.08 52.09 45.45 50.00

MID-RATEDISTRICT

3567 18 51 1609 5 11 526124.39 0.12 0.35 11.00 0.03 0.08 35.9867.80 0.34 0.97 30.58 0.10 0.21 37.10 36.00 30.54 34.05 22.73 26.19

LOW-RATEDISTRICT

2027 9 19 655 7 10 272713.86 0.06 0.13 4.48 0.05 0.07 18.6574.33 0.33 0.70 24.02 0.26 0.37 21.08 18.00 11.38 13.86 31.82 23.81

TOTAL 9615 50 167 4726 22 42 14,62265.76 0.34 1.14 32.32 0.15 0.29 100.00

Table K-6. Crashes by Degree of Curve.District Degree of Curve

FrequencyPercent


Unknown No Curve 0.1 to 1.9 2.0 to 3.9 4.0 to 5.9 6.0 to 7.9 8.0 to 9.9

HIGH-RATE

DISTRICT

249 3663 333 705 618 342 1791.70 25.05 2.28 4.82 4.23 2.34 1.223.75 55.22 5.02 10.63 9.32 5.16 2.70


251 3242 263 448 422 180 841.72 22.17 1.80 3.06 2.89 1.23 0.574.77 61.62 5.00 8.52 8.02 3.42 1.60

41.76 36.62 37.84 32.37 35.31 31.80 29.58LOW-RATE

DISTRICT

101 1948 99 231 155 44 210.69 13.32 0.68 1.58 1.06 0.30 0.143.70 71.43 3.63 8.47 5.68 1.61 0.77

16.81 22.00 14.24 16.69 12.97 7.77 7.39TOTAL 601 8853 695 1384 1195 566 284

4.11 60.55 4.75 9.47 8.17 3.87 1.94


K-8

Table K-6. Crashes by Degree of Curve (continued).District Degree of Curve

FrequencyPercent


10.0 to11.9

12.0 to13.9

14.0 to15.9

16.0 to17.9

18.0 andOver

TOTAL 0 to 3.9 4 to 18+

HIGH-RATE

DISTRICT

209 59 61 17 199 6634 1038 16841.43 0.40 0.42 0.12 1.36 45.373.15 0.89 0.92 0.26 3.00 15.65 25.38


180 41 40 12 98 5261 711 10571.23 0.28 0.27 0.08 0.67 35.983.42 0.78 0.76 0.23 1.86 13.51 20.09

39.47 38.32 36.36 38.71 28.82LOW-RATE

DISTRICT

67 7 9 2 43 2727 330 3480.46 0.05 0.06 0.01 0.29 18.652.46 0.26 0.33 0.07 1.58 12.10 12.76

14.69 6.54 8.18 6.45 12.65TOTAL 456 107 110 31 340 14,622

3.12 0.73 0.75 0.21 2.33 100.00

Table K-7. Crashes by Weather.District Weather

FrequencyPercent


Clear(Cloudy)

Raining Snowing Fog

HIGH-RATEDISTRICT

5930 520 10 16140.56 3.56 0.07 1.1089.39 7.84 0.15 2.4345.43 49.20 17.54 40.76

MID-RATE DISTRICT 4692 364 10 17432.09 2.49 0.07 1.1989.18 6.92 0.19 3.3135.94 34.44 17.54 44.05

LOW-RATE DISTRICT 2432 173 37 6016.63 1.18 0.25 0.4189.18 6.34 1.36 2.2018.63 16.37 64.91 15.19

TOTAL 13,054 1057 57 39589.28 7.23 0.39 2.70


K-9

Table K-7. Crashes by Weather (continued).District Weather

FrequencyPercent


BlowingDust

Smoke Other Sleeting TOTAL

HIGH-RATEDISTRICT

0 2 1 10 66340.00 0.01 0.01 0.07 45.370.00 0.03 0.02 0.15 0.00 40.00 9.09 26.32

MID-RATEDISTRICT

1 1 4 15 52610.01 0.01 0.03 0.10 35.980.02 0.02 0.08 0.29

20.00 20.00 36.36 39.47 LOW-RATEDISTRICT

4 2 6 13 27270.03 0.01 0.04 0.09 18.650.15 0.07 0.22 0.48

80.00 40.00 54.55 34.21 TOTAL 5 5 11 38 14,622

0.03 0.03 0.08 0.26 100.00

Table K-8. Crashes by Surface Condition.District Surface Condition

FrequencyPercent


Dry Wet Muddy Snowy Icy TOTAL

HIGH-RATE DISTRICT 5655 949 2 28 0 663438.67 6.49 0.01 0.19 0.00 45.3785.24 14.31 0.03 0.42 0.00 45.10 49.71 28.57 16.77

MID-RATE DISTRICT 4527 684 3 47 0 526130.96 4.68 0.02 0.32 0.00 35.9886.05 13.00 0.06 0.89 0.00 36.10 35.83 42.86 28.14

LOW-RATE DISTRICT 2357 276 2 92 0 272716.12 1.89 0.01 0.63 0.00 18.6586.43 10.12 0.07 3.37 0.00 18.80 14.46 28.57 55.09

TOTAL 12,539 1909 7 167 0 14,62285.75 13.06 0.05 1.14 0.00 100.00


K-10

Table K-9. Crashes by Light Condition.District Light Condition

FrequencyPercent


Daylight Dawn Dark NotLighted

DarkLighted

Dusk TOTAL

HIGH-RATE DISTRICT 3828 123 2383 171 129 663426.18 0.84 16.30 1.17 0.88 45.3757.70 1.85 35.92 2.58 1.94 45.91 45.72 44.59 43.96 45.74

MID-RATE DISTRICT 2919 90 2000 144 108 526119.96 0.62 13.68 0.98 0.74 35.9855.48 1.71 38.02 2.74 2.05 35.01 33.46 37.43 37.02 38.30

LOW-RATE DISTRICT 1591 56 961 74 45 272710.88 0.38 6.57 0.51 0.31 18.6558.34 2.05 35.24 2.71 1.65 19.08 20.82 17.98 19.02 15.96

TOTAL 8338 269 5344 389 282 14,62257.02 1.84 36.55 2.66 1.93 100.00

Table K-10. Crashes by Month.District Month

FrequencyPercent


January February March April May June July

HIGH-RATEDISTRICT

466 455 514 555 598 495 6283.19 3.11 3.52 3.80 4.09 3.39 4.297.02 6.86 7.75 8.37 9.01 7.46 9.47


385 353 429 421 489 472 4892.63 2.41 2.93 2.88 3.34 3.23 3.347.32 6.71 8.15 8.00 9.29 8.97 9.29

36.39 35.87 36.89 35.17 36.68 39.60 35.51LOW-RATEDISTRICT

207 176 220 221 246 225 2601.42 1.20 1.50 1.51 1.68 1.54 1.787.59 6.45 8.07 8.10 9.02 8.25 9.53

19.57 17.89 18.92 18.46 18.45 18.88 18.88TOTAL 1058 984 1163 1197 1333 1192 1377

7.24 6.73 7.95 8.19 9.12 8.15 9.42


K-11

Table K-10. Crashes by Month (continued).District Month

FrequencyPercent


August September October November December TOTAL

HIGH-RATEDISTRICT

571 553 605 588 606 66343.91 3.78 4.14 4.02 4.14 45.378.61 8.34 9.12 8.86 9.13


420 440 485 445 433 52612.87 3.01 3.32 3.04 2.96 35.987.98 8.36 9.22 8.46 8.23


228 205 241 254 244 27271.56 1.40 1.65 1.74 1.67 18.658.36 7.52 8.84 9.31 8.95

18.70 17.11 18.11 19.74 19.02TOTAL 1219 1198 1331 1287 1283 14,622

8.34 8.19 9.10 8.80 8.77 100.00

Table K-11. Crashes by Day of Week.District Day of Week

FrequencyPercent


Sunday Monday Tuesday Wednesday Thursday Friday Saturday TOTAL

HIGH-RATEDISTRICT

1090 781 740 855 837 1111 1220 66347.45 5.34 5.06 5.85 5.72 7.60 8.34 45.37

16.43 11.77 11.15 12.89 12.62 16.75 18.39 44.51 43.10 44.77 47.79 45.29 47.89 44.35

MID-RATEDISTRICT

902 635 614 599 681 809 1021 52616.17 4.34 4.20 4.10 4.66 5.53 6.98 35.98

17.15 12.07 11.67 11.39 12.94 15.38 19.41 36.83 35.04 37.14 33.48 36.85 34.87 37.11

LOW-RATEDISTRICT

457 396 299 335 330 400 510 27273.13 2.71 2.04 2.29 2.26 2.74 3.49 18.65

16.76 14.52 10.96 12.28 12.10 14.67 18.70 18.66 21.85 18.09 18.73 17.86 17.24 18.54

TOTAL 2449 1812 1653 1789 1848 2320 2751 14,62216.75 12.39 11.30 12.23 12.64 15.87 18.81 100.00


K-12

Table K-12. Crashes by Time.District Time

FrequencyPercent


Midnight -12:59 AM

1-1:59 AM 2-2:59 AM 3-3:59 AM 4-4:59 AM 5-5:59 AM

HIGH-RATEDISTRICT

214 193 176 112 107 1461.46 1.32 1.20 0.77 0.73 1.003.23 2.91 2.65 1.69 1.61 2.20


198 201 164 110 102 1141.35 1.37 1.12 0.75 0.70 0.783.76 3.82 3.12 2.09 1.94 2.17


100 80 75 66 63 740.68 0.55 0.51 0.45 0.43 0.513.67 2.93 2.75 2.42 2.31 2.71

19.53 16.88 18.07 22.92 23.16 22.16TOTAL 512 474 415 288 272 334

3.50 3.24 2.84 1.97 1.86 2.28

Table K-12. Crashes by Time (continued).District Time

FrequencyPercent


6-6:59 AM 7-7:59 AM 8-8:59 AM 9-9:59 AM 10-10:59 AM 11-11:59 AM

HIGH-RATEDISTRICT

202 344 226 202 248 2601.38 2.35 1.55 1.38 1.70 1.783.04 5.19 3.41 3.04 3.74 3.92


171 244 181 151 166 2231.17 1.67 1.24 1.03 1.14 1.533.25 4.64 3.44 2.87 3.16 4.24


96 123 114 94 96 1220.66 0.84 0.78 0.64 0.66 0.833.52 4.51 4.18 3.45 3.52 4.47

20.47 17.30 21.88 21.03 18.82 20.17TOTAL 469 711 521 447 510 605

3.21 4.86 3.56 3.06 3.49 4.14


K-13


FrequencyPercent


NOON-12:59PM

1-1:59 PM 2-2:59 PM 3-3:59 PM 4-4:59 PM 5-5:59 PM

HIGH-RATEDISTRICT

274 291 313 440 432 4291.87 1.99 2.14 3.01 2.95 2.934.13 4.39 4.72 6.63 6.51 6.47


211 242 229 320 297 3451.44 1.66 1.57 2.19 2.03 2.364.01 4.60 4.35 6.08 5.65 6.56


119 113 141 144 172 1640.81 0.77 0.96 0.98 1.18 1.124.36 4.14 5.17 5.28 6.31 6.01

19.70 17.49 20.64 15.93 19.09 17.48TOTAL 604 646 683 904 901 938

4.13 4.42 4.67 6.18 6.16 6.41


FrequencyPercent


6-6:59 PM

7-7:59 PM

8-8:59 PM

9-9:59 PM

10-10:59PM

11-11:59PM

TOTAL

HIGH-RATEDISTRICT

441 321 345 351 303 264 66343.02 2.20 2.36 2.40 2.07 1.81 45.376.65 4.84 5.20 5.29 4.57 3.98

47.22 43.20 47.46 47.30 44.96 46.48 MID-RATEDISTRICT

333 279 277 241 248 214 52612.28 1.91 1.89 1.65 1.70 1.46 35.986.33 5.30 5.27 4.58 4.71 4.07

35.65 37.55 38.10 32.48 36.80 37.68 LOW-RATEDISTRICT

160 143 105 150 123 90 27271.09 0.98 0.72 1.03 0.84 0.62 18.655.87 5.24 3.85 5.50 4.51 3.30

17.13 19.25 14.44 20.22 18.25 15.85 TOTAL 934 743 727 742 674 568 14,622

6.39 5.08 4.97 5.07 4.61 3.88 100.00


K-14

Table K-13. Crashes by Vehicle Movement.District Vehicle Movements / Manner of Collision

FrequencyPercent


Single MotorVehicle Straight

Single MotorVehicle

Right Turn

SingleMotor

Vehicle Left Turn

Single MotorVehicle Backing

Single MotorVehicleOther

Angle BothStraight

HIGH-RATEDISTRICT

4427 26 36 5 1 59230.28 0.18 0.25 0.03 0.01 4.0566.73 0.39 0.54 0.08 0.02 8.9244.76 36.62 36.73 45.45 33.33 44.78

MID-RATEDISTRICT

3515 23 40 4 2 52724.04 0.16 0.27 0.03 0.01 3.6066.81 0.44 0.76 0.08 0.04 10.0235.54 32.39 40.82 36.36 66.67 39.86

LOW-RATEDISTRICT

1949 22 22 2 0 20313.33 0.15 0.15 0.01 0.00 1.3971.47 0.81 0.81 0.07 0.00 7.4419.70 30.99 22.45 18.18 0.00 15.36

TOTAL 9891 71 98 11 3 132267.64 0.49 0.67 0.08 0.02 9.04

Table K-13. Crashes by Vehicle Movement (continued).District Vehicle Movements / Manner of Collision

FrequencyPercent

Row Percent ColPercent

Angle 1Straight 2Backing

Angle 1Straight 2Stopped

Angle 1Straight 2

Right Turn

Angle 1Straight 2Left Turn

Angle BothRight Turn

Angle 1Right Turn

2 LeftTurn

HIGH-RATEDISTRICT

19 17 28 178 0 00.13 0.12 0.19 1.22 0.00 0.000.29 0.26 0.42 2.68 0.00 0.00


9 22 21 124 0 10.06 0.15 0.14 0.85 0.00 0.010.17 0.42 0.40 2.36 0.00 0.02


0 7 11 49 0 00.00 0.05 0.08 0.34 0.00 0.000.00 0.26 0.40 1.80 0.00 0.000.00 15.22 18.33 13.96 0.00

TOTAL 28 46 60 351 0 10.19 0.31 0.41 2.40 0.00 0.01


K-15


FrequencyPercent


Angle 1 RightTurn

2 Stopped

Angle BothLeft Turn

Angle 1 LeftTurn 2

Stopped

SameDirection

Both StraightRear End

SameDirection

Both StraightSideswipe

SameDirection 1Straight 2Stopped

HIGH-RATEDISTRICT

2 0 2 173 29 1740.01 0.00 0.01 1.18 0.20 1.190.03 0.00 0.03 2.61 0.44 2.62


3 0 2 128 10 1090.02 0.00 0.01 0.88 0.07 0.750.06 0.00 0.04 2.43 0.19 2.07


1 0 0 73 10 540.01 0.00 0.00 0.50 0.07 0.370.04 0.00 0.00 2.68 0.37 1.98

16.67 0.00 19.52 20.41 16.02TOTAL 6 0 4 374 49 337

0.04 0.00 0.03 2.56 0.34 2.30


FrequencyPercent


SameDirection 1Straight 2

Right Turn

SameDirection 1Straight 2Left Turn

SameDirection

Both RightTurn

SameDirection 1Right Turn2 Left Turn

SameDirection 1Right Turn 2 Stopped

SameDirectionBoth Left

Turn

HIGH-RATEDISTRICT

42 220 0 0 1 00.29 1.50 0.00 0.00 0.01 0.000.63 3.32 0.00 0.00 0.02 0.00

39.25 41.43 0.00 100.00 MID-RATEDISTRICT

36 207 2 0 0 00.25 1.42 0.01 0.00 0.00 0.000.68 3.93 0.04 0.00 0.00 0.00

33.64 38.98 100.00 0.00 LOW-RATEDISTRICT

29 104 0 0 0 00.20 0.71 0.00 0.00 0.00 0.001.06 3.81 0.00 0.00 0.00 0.00

27.10 19.59 0.00 0.00 TOTAL 107 531 2 0 1 0

0.73 3.63 0.01 0.00 0.01 0.00


K-16


FrequencyPercent


SameDirection 1Left Turn 2

Stopped

OppositeDirections

BothStraight

OppositeDirections 1Straight 2Backing

OppositeDirections 1Straight 2Stopped

OppositeDirections 1

Straight2 Right Turn

OppositeDirections 1Straight 2Left Turn

HIGH-RATEDISTRICT

0 378 2 9 2 2570.00 2.59 0.01 0.06 0.01 1.760.00 5.70 0.03 0.14 0.03 3.87


0 253 0 8 1 2030.00 1.73 0.00 0.05 0.01 1.390.00 4.81 0.00 0.15 0.02 3.86


0 115 1 3 1 650.00 0.79 0.01 0.02 0.01 0.440.00 4.22 0.04 0.11 0.04 2.38

15.42 33.33 15.00 25.00 12.38TOTAL 0 746 3 20 4 525

0.00 5.10 0.02 0.14 0.03 3.59


FrequencyPercent


OppositeDirections 1Backing 2Stopped

OppositeDirections 1Right Turn 2 Left Turn

OppositeDirections 1Right Turn 2 Stopped

OppositeDirections

Both Left Turn

OppositeDirections 1Left Turn2 Stopped

Other 1Straight2 Parked

HIGH-RATEDISTRICT

0 1 0 0 0 70.00 0.01 0.00 0.00 0.00 0.050.00 0.02 0.00 0.00 0.00 0.11

50.00 0.00 35.00MID-RATEDISTRICT

0 1 1 1 0 80.00 0.01 0.01 0.01 0.00 0.050.00 0.02 0.02 0.02 0.00 0.15

50.00 100.00 40.00LOW-RATEDISTRICT

0 0 0 0 0 50.00 0.00 0.00 0.00 0.00 0.030.00 0.00 0.00 0.00 0.00 0.18

0.00 0.00 25.00TOTAL 0 2 1 1 0 20

0.00 0.01 0.01 0.01 0.00 0.14


K-17


FrequencyPercent


Other 1RightTurn

2 Parked

Other 1Left Turn2 Parked

Other 1Parked

2 Stopped

OtherBoth

Parked

OtherBoth

Backing

Other AllOthers

TOTAL

HIGH-RATE DISTRICT 0 0 0 0 0 6 66340.00 0.00 0.00 0.00 0.00 0.04 45.370.00 0.00 0.00 0.00 0.00 0.09

85.71 MID-RATE DISTRICT 0 0 0 0 0 0 5261

0.00 0.00 0.00 0.00 0.00 0.00 35.980.00 0.00 0.00 0.00 0.00 0.00

0.00 LOW-RATE DISTRICT 0 0 0 0 0 1 2727

0.00 0.00 0.00 0.00 0.00 0.01 18.650.00 0.00 0.00 0.00 0.00 0.04

14.29 TOTAL 0 0 0 0 0 7 14,622

0.00 0.00 0.00 0.00 0.00 0.05 100.00

Table K-14. Crashes by Other Factor.District Other Factor

FrequencyPercent


No Code Applicable

Lost Control,Skidded

PassengerInterfered

AttentionDiverted

ObjectProjecting

fromVehicle

FootSlipped

Off Brake

HIGH-RATEDISTRICT

4688 32 18 238 1 732.06 0.22 0.12 1.63 0.01 0.0570.67 0.48 0.27 3.59 0.02 0.1145.38 23.02 41.86 39.87 25.00 100.00

MID-RATEDISTRICT

3691 42 15 226 3 025.24 0.29 0.10 1.55 0.02 0.0070.16 0.80 0.29 4.30 0.06 0.0035.73 30.22 34.88 37.86 75.00 0.00

LOW-RATEDISTRICT

1952 65 10 133 0 013.35 0.44 0.07 0.91 0.00 0.0071.58 2.38 0.37 4.88 0.00 0.0018.89 46.76 23.26 22.28 0.00 0.00

TOTAL 10,331 139 43 597 4 770.65 0.95 0.29 4.08 0.03 0.05


K-18

Table K-14. Crashes by Other Factor (continued).District Other Factor

FrequencyPercent


Gusty Winds

Vehicle Passing On Left

VehiclePassing

On Right

VehicleChanging

Lane

ImproperlyParkedVehicle

VehicleForward

fromParking

HIGH-RATEDISTRICT

2 131 11 8 8 60.01 0.90 0.08 0.05 0.05 0.040.03 1.97 0.17 0.12 0.12 0.09


6 109 6 4 5 80.04 0.75 0.04 0.03 0.03 0.050.11 2.07 0.11 0.08 0.10 0.15


8 62 7 6 3 70.05 0.42 0.05 0.04 0.02 0.050.29 2.27 0.26 0.22 0.11 0.26

50.00 20.53 29.17 33.33 18.75 33.33TOTAL 16 302 24 18 16 21

0.11 2.07 0.16 0.12 0.11 0.14


FrequencyPercent


VehicleBackward

fromParking

VehicleEntering

Driveway

VehicleLeaving

Driveway

VisionObstructed by

StandingVehicle

VisionObstructedby Moving

Vehicle

VisionObstructedby Embank-

mentHIGH-RATEDISTRICT

0 230 112 5 6 10.00 1.57 0.77 0.03 0.04 0.010.00 3.47 1.69 0.08 0.09 0.02


0 160 70 7 6 00.00 1.09 0.48 0.05 0.04 0.000.00 3.04 1.33 0.13 0.11 0.00


0 56 28 0 1 00.00 0.38 0.19 0.00 0.01 0.000.00 2.05 1.03 0.00 0.04 0.00

12.56 13.33 0.00 7.69 0.00TOTAL 0 446 210 12 13 1

0.00 3.05 1.44 0.08 0.09 0.01


K-19


FrequencyPercent


VisionObstructed byCommercial

Sign

VisionObstructed byHighway Sign

VisionObstructed by

Glare

VisionObstructed by

Hillcrest

VisionObstructed by

Trees

VisionObstructed by

Other

HIGH-RATEDISTRICT

1 0 42 10 1 610.01 0.00 0.29 0.07 0.01 0.420.02 0.00 0.63 0.15 0.02 0.92


0 1 30 3 3 690.00 0.01 0.21 0.02 0.02 0.470.00 0.02 0.57 0.06 0.06 1.310.00 50.00 37.04 20.00 60.00 44.23

LOW-RATEDISTRICT

0 1 9 2 1 260.00 0.01 0.06 0.01 0.01 0.180.00 0.04 0.33 0.07 0.04 0.950.00 50.00 11.11 13.33 20.00 16.67

TOTAL 1 2 81 15 5 1560.01 0.01 0.55 0.10 0.03 1.07


FrequencyPercent


Swerved to ChangeLanes

Swerved, Not Specified

Swerved,Surface orVisibility

Swerved,TrafficControl

Swerved, Pedestrianor Cyclist

Swerved,Animal

HIGH-RATEDISTRICT

9 84 1 0 2 2650.06 0.57 0.01 0.00 0.01 1.810.14 1.27 0.02 0.00 0.03 3.99


13 66 1 2 3 2090.09 0.45 0.01 0.01 0.02 1.430.25 1.25 0.02 0.04 0.06 3.97


3 22 0 1 0 1130.02 0.15 0.00 0.01 0.00 0.770.11 0.81 0.00 0.04 0.00 4.14

12.00 12.79 0.00 33.33 0.00 19.25TOTAL 25 172 2 3 5 587

0.17 1.18 0.01 0.02 0.03 4.01


K-20


FrequencyPercent


Swerved,Object in

Road

Swerved, SlowVehicle

Swerved,Vehicle Entering

Road

Swerved,AvoidingVehicle in

Wrong Lane

Swerved, AvoidingPreviousAccident

Slowed, NotSpecified

HIGH-RATEDISTRICT

19 31 33 101 0 370.13 0.21 0.23 0.69 0.00 0.250.29 0.47 0.50 1.52 0.00 0.56


11 19 40 59 3 450.08 0.13 0.27 0.40 0.02 0.310.21 0.36 0.76 1.12 0.06 0.86


7 8 8 21 0 170.05 0.05 0.05 0.14 0.00 0.120.26 0.29 0.29 0.77 0.00 0.62

18.92 13.79 9.88 11.60 0.00 17.17TOTAL 37 58 81 181 3 99

0.25 0.40 0.55 1.24 0.02 0.68


FrequencyPercent


Slowed,Surface orVisibility

Slowed, Traffic Control

Slowed, Pedestrianor Cyclist

Slowed, Animal

Slowed,Object in

Road

Slowed,Slow

Vehicle

HIGH-RATEDISTRICT

2 16 1 7 3 460.01 0.11 0.01 0.05 0.02 0.310.03 0.24 0.02 0.11 0.05 0.69


1 26 0 4 0 220.01 0.18 0.00 0.03 0.00 0.150.02 0.49 0.00 0.08 0.00 0.42


1 7 0 3 0 70.01 0.05 0.00 0.02 0.00 0.050.04 0.26 0.00 0.11 0.00 0.26

25.00 14.29 0.00 21.43 0.00 9.33TOTAL 4 49 1 14 3 75

0.03 0.34 0.01 0.10 0.02 0.51


K-21


FrequencyPercent


Slowed,Vehicle Entering

Road

Slowed,Vehicle in

Wrong Lane

Slowed,AvoidingPreviousAccident

Slowed,TurningRight

Slowed,Turning

Left

SchoolBus Related

Accident

HIGH-RATEDISTRICT

5 1 0 23 144 240.03 0.01 0.00 0.16 0.98 0.160.08 0.02 0.00 0.35 2.17 0.36


3 2 1 12 79 160.02 0.01 0.01 0.08 0.54 0.110.06 0.04 0.02 0.23 1.50 0.30


1 2 0 10 42 60.01 0.01 0.00 0.07 0.29 0.040.04 0.07 0.00 0.37 1.54 0.22

11.11 40.00 0.00 22.22 15.85 13.04TOTAL 9 5 1 45 265 46

0.06 0.03 0.01 0.31 1.81 0.31


FrequencyPercent


HighwayConstructionUnrelated toConstruction

HighwayConstruction

Related toConstruction

OtherConstruction

AreaUnrelated toConstruction

OtherConstructionArea Related

toConstruction

Accident onBeach

TOTAL

HIGH-RATEDISTRICT

125 34 0 2 0 00.85 0.23 0.00 0.01 0.00 0.001.88 0.51 0.00 0.03 0.00 0.0041.39 40.00 40.00

MID-RATEDISTRICT

126 32 0 2 0 00.86 0.22 0.00 0.01 0.00 0.002.39 0.61 0.00 0.04 0.00 0.0041.72 37.65 40.00

LOW-RATEDISTRICT

51 19 0 1 0 00.35 0.13 0.00 0.01 0.00 0.001.87 0.70 0.00 0.04 0.00 0.0016.89 22.35 20.00

TOTAL 302 85 0 5 0 02.07 0.58 0.00 0.03 0.00 0.00

Documents

Form DOT F 1700 · KAB Crashes 4824 39.6 44,606 25.9 Intersection 1734 14.2 41,112 23.8 Intersection-Related 1331 10.9 32,798 19.0 Driveway Access Related 1092 9.0 16,296 9.4 Non-Intersection